The Pharmacology of NSAIDs
Rationale for the Use of NSAIDs in OA
In case of inflammation, NSAIDs can switch off peripheral sensitization by inhibiting a relevant amount of PGs. Thus, their use as first-line therapy aimed at treating inflammatory nociceptive pain is virtually always appropriate; however, the feasibility of such strategy depends on the condition of the patient.
Conversely, the use of paracetamol (very common in OA although it is not an NSAID) is inappropriate in inflammatory pain, since it is a weak inhibitor of COX-1 and a very weak inhibitor of COX-2 and, as such, it does not interfere with peripheral sensitization. In addition, in tissues with inflammation, the free radicals inactivate paracetamol, abolishing any action on COX-2 [50]. In line with these observations, its analgesic effect cannot depend on COX inhibition. Paracetamol is actually a central analgesic with multiple effects, the main one being the stimulation of the endogenous cannabinoid system [51]. Hence, paracetamol has a lower efficacy than NSAIDs in reducing inflammatory pain [21] and its central analgesic efficacy is also lower than that of opioids.
Are All NSAIDs Equal?
All NSAIDs have an inhibitory activity on COX-1 and COX-2 but there are several differences among NSAIDs (for details on the mechanisms of action of the most common NSAIDs refer to [52,53,54]) that impact on their efficacy and safety [54]. These include:
-
Chemical similarity
-
COX isoform selectivity and potency [55]. NSAIDs comprise non-selective drugs, such as ibuprofen and naproxen, and selective COX-2 inhibitors (i.e. coxibs), such as etoricoxib and celecoxib. Potency is not a synonym of selectivity and cannot be used to predict dosages: a drug is potent if it inhibits 50% of available COX-1 and COX-2 at low dose. For example, etoricoxib is a selective inhibitor of COX-2 but it is less potent than diclofenac which, together with ketorolac (which has no indication in OA treatment), is the most potent inhibitor of COX-2 [55].
Importantly, the kinetics of COX-1 and COX-2 inhibition are different (non-linear and linear, respectively [56]).
In clinical practice, to achieve a significant anti-thrombotic effect through the blockade of thromboxane A synthesis, 95–97% of platelet COX-1 must be inhibited. If 90% of the enzyme is blocked, no anti-thrombotic effect occurs. The only NSAID able to inhibit 95% of COX-1 is acetylsalicylic acid (ASA), which irreversibly blocks the enzyme and, if administered at the dose of 100 mg per day every day, maintains this level of inhibition. No other NSAID is able to produce this effect, with the exception of naproxen but at non-conventional doses and regimens. Thus, it is not completely true that NSAIDs alone interfere with platelet function. Certainly, there are COX-1-independent antiplatelet effects that may play a role. NSAIDs induce mostly GI bleeding as they block COX-1 in surface epithelial cells. Given in concomitance with other drugs, such as the selective serotonin reuptake inhibitors, the antiplatelet effects of NSAIDs are enhanced.
-
Plasma half-life. This feature impacts on the occurrence of AEs. Indeed, the NSAIDs inhibiting gastric COX-1 for a longer time are more harmful for the stomach. For example, piroxicam and diclofenac have a half-life of about 60 and 1 h, respectively, but the latter is a more potent inhibitor of COX-1 and is associated to a relative risk of gastric bleeding of 3.61 compared to 8.00 for piroxicam [57].
-
Interference with ASA. Not all NSAIDs interfere with the cardioprotective effects of ASA. Diclofenac, ketorolac and etoricoxib do not, but they are the most potent inhibitors of COX-2 and, thus, of the endothelial prostacyclin (PGI2) production. Upon ASA treatment, the levels of thromboxane A drop and only the endogenous PGI2 remain, leading to a “thrombotic equilibrium.” If COX-2 is blocked, the equilibrium is impaired again. Ibuprofen, but not etoricoxib or diclofenac, seems to interfere with the capability of ASA to irreversibly acetylate platelet COX-1. This might reduce the protective effect of ASA against the risk of atherothrombotic events. Notably, combining ASA (required to prevent CV events) with a coxib may enhance the protective effect of COX-2 inhibition toward the gastric mucosal and prolong the time to recover from gastric mucosal injury [58].
According to the pharmacologist, patients on ASA must not take any NSAID. In particular cases, such as of a gout flare or of a renal colic, they may take such therapy for 1–2 days.
-
Penetration into the synovial liquid. Not all NSAIDs adequately penetrate into the synovial liquid (e.g. ibuprofen does not while diclofenac does), so even in the case of a short half-life, the higher the absorption at the synovial site, the longer the pharmacological effect [59].
-
Passage through the blood–brain barrier. This aspect related to the central action of NSAIDs may be of interest when selecting the most appropriate drug. Some compounds, such as diclofenac, pass through the barrier and reach the spinal cord, where the PGs produced by neurons and astroglia play a role in central sensitization. Therefore, at this site, inhibition of COX-1 and COX-2 adds to the peripheral effect (possible synergism) so that the analgesic activity resulting from the anti-inflammatory action adds to a central analgesic effect occurring when high drug doses reach the central nervous system.
Factors Influencing the Individual Response to NSAIDs
Several players affect the inter-patient variability observed in the response to NSAID therapy:
-
(1)
Genetic variations in the enzymes that metabolize NSAIDs (cytochrome P450 2C9 [CYP2C9] in many cases) and COXs.
-
(2)
The microbiota, for its capability to inactivate drugs. However, data in this regard are scarce.
-
(3)
The possibility of phenotyping OA (e.g. coxarthrosis vs. gonarthrosis), which is rather concrete [35] and may help to decide if and how to use an NSAID therapy—the choice should rely on the evidence from head-to-head comparisons or network meta-analyses [19,20,21].
-
(4)
Gender, which is responsible for relevant differences in the incidence, prevalence and prognosis of several immunoinflammatory diseases. Pre-clinical studies have demonstrated that the molecular mechanisms of inflammation and pain may differ between men and women. All of these differences provide a plausible background to understand why women use more NSAIDs than men. However, the pharmacological mechanisms underlying the gender-driven NSAID responses remain elusive [60].
By When Should We Expect the Response to NSAID treatment?
Usually, the maximum peak plasma concentration is reached within 2–3 h of administration, but the efficacy also depends on other factors (e.g. plasma protein binding and tissue distribution with particular regard to the inflammatory osteoarticular tissue). The rapid effect is pain reduction, which is achieved also through the central activity of NSAIDs; the delayed effect is the reduction of inflammation and thus the rise of the threshold; the variable effect is the improvement in disability. The analgesic effect occurs within about 1 week and the full anti-inflammatory effect is often achieved in 3 weeks (which questions the 3-day test validity, as the specificity is very low) [61]. A recent study has shown that the NSAID-induced improvement in pain and function peaks at 2 weeks and starts to decline by 8 weeks, while minor CV and GI AEs occur as early as 4 weeks after the initiation of NSAID treatment [62].
What is the Adequate Duration of NSAID Therapy?
In general, NSAIDs should be used for the shortest duration possible and at the lowest dose that guarantees both inflammation reduction and physical function improvement, as established in efficacy studies [21, 61]. Therapy duration must be tailored to the patient profile [61]. Usually, the treatment duration is at least 7–10 days, taking into account the time required to achieve both the analgesic and full anti-inflammatory effects [61]. If at the end of the 3-week period no result has occurred, a switch to another agent should be attempted [61].
Monotherapy or Combination Therapy?
It is possible to combine NSAIDs with central analgesics, such as paracetamol and opioids. By targeting different mechanisms, such combinations permit the dose to be limited, thus reducing the risk of AEs. In contrast, the combination of NSAIDs with steroids should be avoided: in fact, although they are very effective against inflammation and cause only marginal gastric erosion in subjects without risk factors, these drugs delay the healing of possible microulcers, highly enhancing the NSAID-induced gastric erosion. In this context, the number of administrations plays a central role.
General Considerations on the Different Formulations
Oral intake through the direct contact between drugs and the GI tract mucosa increases the likelihood of topical damage until absorption. Topical formulations are usually preferred over systemic treatments for safety reasons, such as in patients aged > 75 years [4, 5]. In patients with comorbidities, to favor compliance, formulations relying on one or few administrations (e.g. modified release) should be considered.
Practical Indications
-
(1)
Avoid the use of paracetamol in case of inflammatory pain.
-
(2)
NSAIDs should be used for the shortest duration and at the lowest dose that guarantees the effect on inflammation and improvement in physical function.
-
(3)
Define therapy duration based on the patient profile and avoid the on-demand use of NSAIDs: in the case of inflammatory pain, therapy must be administered for at least 10 days to achieve analgesia and for 3 weeks to achieve the full anti-inflammatory effect.
-
(4)
It is possible to combine NSAIDs with central analgesics such as paracetamol and opioids.
-
(5)
Avoid the combination of NSAIDs with steroids.
-
(6)
Consider formulations relying on one or few administrations to improve adherence.
Making Sense of NSAID Therapy: The Specialists’ Point of View
In clinical practice, in-depth knowledge of each NSAID’s efficacy and safety profile, together with the patient characteristics, is critical to define the benefit/risk balance of each compound for a specific individual and drive the therapeutic choice.
Table 2 summarizes the considerations made by the GP, the pharmacologist and the pain therapist of the multidisciplinary panel.
Table 2 Considerations driving the choice of therapy according to the general practitioner, the pharmacologist and the pain therapist The Safety Profile of NSAIDs
The main AEs that may occur upon NSAID therapy are illustrated in Fig. 2.
NSAIDs and CV Risk
NSAID-Related CV AEs
The CV safety of NSAIDs is a very controversial matter. Following the observation that NSAIDs could increase the risk of CV events at therapeutic doses or higher, in 2005 the U.S. Food and Drug Administration added a black box warning to their use [63], while the European Medicines Agency decided to contraindicate coxibs (but not tNSAIDs [64]) in patients with coronary heart disease or stroke and to advise those at risk for coronary heart disease to use these agents with caution [65].
The possible mechanisms proposed to explain CV complications include (1) the unbalance between the vasodilator effect of PGI2 and PGE2 in favor of vasoconstriction by thromboxane A2 in the endothelium, which results in a prothrombotic effect; and (2) sodium and water retention promoted by COX inhibition, which worsens heart failure, hypertension and ventricular remodeling.
The Coxib and Traditional NSAID Trialists’ (CNT) Collaboration meta-analysis is the largest meta-analysis on NSAID safety, based on 639 RCTs in which tNSAIDs/coxibs were used for long periods [26]. It investigated the vascular effects of coxibs (celecoxib, etoricoxib and lumiracoxib) and high-dose tNSAIDs (diclofenac, ibuprofen and naproxen) in older patients with rheumatic diseases [26]. Coxibs, diclofenac and ibuprofen displayed a similar relative risk for CV events (range 1.37–2.49), whereas naproxen did not seem to increase it (range 0.39–1.87). Coxibs, diclofenac and ibuprofen also displayed a comparable annual absolute risk for major vascular events, which varied according to the baseline predicted risk: in low-risk subjects, the predicted absolute risk of major vascular events was low regardless of the NSAID administered (2 per 1000 in all cases for coxibs, diclofenac and ibuprofen; 0 per 1000 for naproxen); in high-risk patients, the risk increased and was similar for high-dose diclofenac and coxibs (8 per 1000 and 7 per 1000, respectively) and possibly ibuprofen (9 pe 1000), while it seemed to be lower for high-dose naproxen (− 1 per 1000). [26]. A subsequent network meta-analysis found no difference in the risk of major CV events with diclofenac, ibuprofen, naproxen, celecoxib and etoricoxib for the treatment of pain in patients with OA or RA [20].
The PRECISION trial, conducted in subjects with OA or RA at increased CV risk and treated with celecoxib, naproxen and ibuprofen, showed a similar number of CV-related deaths, nonfatal myocardial infarction (MI) or nonfatal stroke among the three groups of NSAIDs, but ibuprofen and naproxen had been used at doses and for periods not in line with guidelines [66].
The absolute risk for CV effects increases to a greater extent in patients with or at risk for active atherosclerotic processes (e.g. with recent bypass surgery, unstable angina or ischemic cerebrovascular events) receiving a COX inhibitor. The excess number of events depends on the underlying risk of the patient, the relative risk of the drug and the duration of the follow-up [58].
A recent meta-analysis of individual patient data in real-world settings [30] has shown that all traditional NSAIDs are associated with an increased risk of AMI, similar to that reported with celecoxib therapy. Using a high daily dose (celecoxib > 200 mg, diclofenac > 100 mg, ibuprofen > 1200 mg, naproxen > 750 mg) for 8–30 days was associated with the greatest risk, which did not increase further beyond the first 30 days. Based on these findings, prescribers should consider weighing the risks and benefits of NSAIDs before selecting the treatment, particularly for higher doses.
In patients with a prior MI, the excess risk of mortality has been estimated to be approximately six deaths per 100 person-years of treatment with a COX-2 inhibitor compared with no NSAID treatment [67]. A Danish large-scale study based on national administrative registers and conducted in healthy individuals demonstrated an increased risk for death/MI in diclofenac and celecoxib users (hazard ratio [95% confidence interval] vs. non-users: 1.63 [1.52–1.76] and 2.01 [1.78–2.27], respectively), which increased in a dose-dependent fashion [68]. Moreover, in low-risk patients, an increased risk of pooled CV events was found with lower doses of diclofenac versus paracetamol, ibuprofen and naproxen (which, however, varied based on the event considered) over 1 month; surprisingly, the relative risk decreased in patients at high CV risk [69], but the explanation remains unclear [24]. Finally, the SOS project, which included millions of Europeans, showed a similar modest increase in CV risk with diclofenac and other NSAIDs, compared to non-use [70]. Nonetheless, bias linked to the design of the available studies do not allow definitive conclusions to be drawn.
A recent study showed that patients on anticoagulant therapy with both vitamin K antagonists and dabigatran should avoid NSAIDs due to a greater risk of hemorrhage, especially GI bleeding, and more frequent complications (such as strokes and embolisms) [71]. The use of NSAIDs has always been discouraged in patients receiving antivitamin K therapy, but this is valid advice also for those who receive dabigatran and likely all direct anticoagulants. No specific data are currently available for rivaroxaban, edoxaban and apixaban.
Finally, it must be pointed out that the concomitant administration of certain NSAIDs weakens the protective CV effects of ASA [72,73,74]. Co-administration of ibuprofen in patients with documented CVD on low-dose ASA therapy significantly increased the risk of all-cause and CV mortality (hazard ratio [HR] 1.93, 95% confidence interval [CI] 1.30–2.87; HR 1.73, 95% CI 1.05–2.84, respectively) compared to ASA alone [72]. No difference was observed when diclofenac or other NSAIDs were used with ASA versus ASA alone [72].
In conclusion, the main determinant of the risk of AEs is the patient profile. A slight increase in CV risk occurs mainly in case of high doses and long-term use, which, however, are not recommended by current guidelines. The only indication to limit the CV risk is to adhere to the recommended dosages and duration and, possibly, undergo cycles of therapy with periodical interruptions.
Practical Indications
-
(1)
In patients with a prior MI, extra caution is needed in the use of NSAIDs/coxibs.
-
(2)
Use only the recommended doses and for the shortest period necessary to control or relieve symptoms.
-
(3)
Monitor renal function and blood pressure in NSAID/coxib users, especially if they present preexisting conditions such as hypertension, renal disease and heart failure [75].
NSAIDs and GI Risk
NSAID-Related GI AEs
Non-steroidal anti-inflammatory drugs-induced GI AEs are various and sometimes severe even though their prevalence is not high [16]. The most frequent GI AEs assocated with NSAID therapy are gastric injuries, which range from subjective manifestations, such as dyspepsia, to ulcers with complications. In elderly patients with arthritis, the incidence of GI intolerability AEs was reported to be significantly lower with celecoxib (16.7%) than with naproxen (29.4%; P < 0.0001), ibuprofen (26.5%; P = 0.0016) and diclofenac (21.0%; P < 0.0001). The discontinuation rate due to these AEs was similar for celecoxib (4.0%) and diclofenac (4.2%; P = 0.75) and significantly lower than for naproxen (8.1%; P < 0.0001) and ibuprofen (7.3%; P < 0.05) [76].
Up to 70% of NSAID users experience minimal mucosal lesions as early as within a few hours of intake [77]; these may indicate gastric mucosa frailty and the tendency to become real ulcers [78]. NSAID-induced ulcer, mainly gastric, is becoming more and more frequent due to the increased use of these drugs, especially in the elderly. As NSAID use and Helicobacter pylori are two independent determinants of ulcer development, they may have additive effects on the ulcer risk in the same subject. Thus, the most recent international guidelines recommend that patients be tested for the presence of the infection and, if present, to eradicate it in those who have to start a prolonged therapy with NSAIDs [79]. The most frequent ulcer complication is bleeding, with a rate ratio (RR) of 1–2% per year. The underlying disease seems to be important: for example, the rate of bleeding is 1.3–2% per year in RA patients and 0.7–10% per year in those with OA [80].
NSAID users may also experience intestinal disorders, including small bowel injuries [77], which may be caused by the mucosal inflammatory pathway triggered by microbiota changes [81].
Liver toxicity events are much less frequent than gastric injuries. Paracetamol used at high doses, at least 4 g per day, may damage the liver [82]. Other studies found that the RR of liver damage defined by hypertransaminasemia was higher for nimesulide (2.2) and sulindac (5) than for diclofenac (1.5) [82].
Most of the patients who develop a serious GI AE while on NSAID therapy are asymptomatic prior to the event [83], particularly the elderly. Among the risk factors for the onset of NSAID-associated ulcer complications (Table 1), advanced age is a primary risk factor for GI events [84]: indeed, NSAID users aged 75–89 years have a twofold higher risk of bleeding (RR 4.1) compared to users aged 60–74 years (RR 2.0) [85]. It is frequent to observe, in the emergency department, elderly patients who use NSAIDs chronically and present severe anemia with hemoglobin levels of 4–5 gr/dl without having ever experienced any dyspeptic symptom. Therefore, physicians must check their patients periodically for the presence of anemia (fecal occult blood test, hematocrit) and symptoms associated with this condition (headache, asthenia, dyspnea, etc.). Conversely, many patients with troublesome symptoms (e.g. epigastric pain and dyspepsia) may have a normal endoscopy at the upper GI tract [86]. As for steroids increasing the risk of complications, it must be pointed out that, when used alone, they do not represent an actual risk for ulcerogenesis [87].
The type of non-selective NSAID impacts on the frequency of GI damage. The results from two epidemiological studies have led to establish a scale of risk for different tNSAIDs (i.e. ibuprofen, diclofenac, naproxen, ketoprofen, indomethacin, piroxicam and azapropazone); azapropazone and piroxicam were associated to the highest risk of gastroduodenal bleeding (odds ratio [OR] 23.4–31.5 and 13.7–18, respectively) and diclofenac and ibuprofen with the lowest (OR 3.9–4.2 and 2.0–2.9, respectively) [88, 89]. Table 3 presents the results from two recent meta-analyses of RCTs that report the rate of risk for bleeding associated with tNSAIDs and coxibs versus placebo [26] and for major GI events associated with tNSAIDs and coxibs versus diclofenac [20]. In particular, the CNT meta-analysis reported that the annual absolute risk of upper GI complications for coxibs, diclofenac, ibuprofen and naproxen depended on the baseline risk [26]. Both in patients at low and high risk, diclofenac and coxibs yielded a similar risk (in low-risk patients: 2 per 1000; in high-risk patients: 6 per 1000, respectively) that was lower than that of ibuprofen and naproxen (in low-risk patients: 4 per 1000; in high-risk patients: 15 and 16 per 1000, respectively), in line with the results from previous epidemiological studies [88, 89].
Table 3 Risk of gastroduodenal bleeding or overall gastrointestinal complications according to the NSAID administered Among the NSAID features that may impact on gastrolesivity, plasma half-life plays a major role. A study conducted in elderly subjects [90] evaluated the presence of gastroduodenal bleeding through the measurement of fecal blood loss and found that it was higher with drugs with a longer plasma half-life, such as naproxen (2.76 ml fecal blood loss) and piroxicam (1.16 ml), compared to diclofenac (0.53 ml), a NSAID with a shorter half-life, and placebo (0.28 ml). Other factors responsible for a different gastrolesive effect among NSAIDs are the level of pK (higher levels increase the toxic effect) and the dosage.
When is it Adequate to Use Proton Pump Inhibitors with NSAIDs in the Prevention of NSAID-Induced Damage?
Non-steroidal anti-inflammatory drug-induced GI damage can be significantly reduced by increasing the gastric pH through the administration of proton pump inhibitors (PPIs), which are the most potent acid inhibitors available. Unlike H2-antagonists that prevent only the onset of duodenal ulcers, PPIs can protect both the stomach, the main site of NSAID-induced damage, and the duodenum [91]. The protective action of PPIs depends on the fact that the weakening of the mechanisms of mucosal defense induced by NSAIDs implies that even a reduced amount of acid, such as in the case of the chronic gastritis that is always associated to gastric ulcers, may be dangerous [92]. A number of important risk factors must be considered due to the need to administer an appropriate prophylactic therapy with PPIs [93] (Table 4).
Table 4 Risk factors in NSAID users requiring prophylaxis with proton pump inhibitors PPIs have to be administered throughout the period of NSAID use; even half the standard dose seems to be sufficient to achieve the benefit [94].
Practical Indications
-
(1)
There are no dietary or behavioral suggestions to prevent or reduce NSAID-induced GI lesions.
-
(2)
When selecting a NSAID, in high-risk patients or in case of prolonged therapy duration, compounds with the lowest risk of GI events should be preferred.
-
(3)
The optimal treatment duration depends on the disease and corresponds to the period of acute symptoms or of functional joint impairment.
-
(4)
Use PPIs in the presence of particular risk factors.
NSAIDs and Renal AEs
NSAID-Induced Renal and Reno-Vascular Events and Risk Factors
At the center stage of the untoward effects of NSAIDs is the inhibition of endogenous or inflammatory renal PGs, a subfamily of eicosanoids. Endogenous eicosanoids fine-tune renal microcirculation and water and electrolyte transport across renal tubules. PGs, such as PGE1 and 2 or PGF2α, control sodium reabsorption and the concentration/dilution mechanism. Likewise, endothelial PGI2 and platelet thromboxane A2 balance each other to control vascular tone in glomeruli and renal arterioles, including the vasa recta. As this counterbalance mechanism is marginal in the normal kidney, NSAID inhibitors of eicosanoid biosynthesis have very modest effects in the healthy kidney and/or younger individuals and are usually well tolerated in persons with normal renal function.
Elderly individuals or patients with chronic kidney disease (CKD) are likely to experience at least some mild AEs, ranging from local edema (e.g. hands, lower limbs, water retention with rapid weight gain) to worsening of glomerular filtration rate (GFR) and/or hyperkalemia. This is usually more frequent in patients with certain comorbidities: in particular, in hypertensive subjects NSAID therapy may lead to intensification of anti-hypertensive regimen [95]. The effects are usually reversible but tend to synergize with other agents affecting renal function, such as anti-hypertensive drugs. In selected circumstances, acute kidney injury (AKI) may occur with severe oligoanuria. A meta-analysis of observational studies [28] found a statistically significant elevated AKI risk in patients treated with indomethacin, piroxicam, ibuprofen, naproxen and sulindac versus non-users, with pooled RRs ranging from 1.58 to 2.11. In all other cases (i.e. diclofenac, meloxicam, and celecoxib), the increase in AKI risk was not significant. Another meta-analysis of observational studies [29] reported that, in the general population, the pooled OR of AKI for ongoing NSAID exposure was 1.73 (95% CI 1.44–2.07) and was higher in older people (OR 2.51, 95% CI 1.52–2.68); in people with CKD, it was 1.63 (95% CI 1.22–2.19) and ranged from 1.12 to 5.25. Notably, the risk was higher for NSAIDs with no COX-2 selectivity (OR 1.84, 95% CI 1.54–2.19) and decreased with increasing COX-2 selectivity (≥ 5-fold, OR 1.41, 95% CI 1.07–1.87).
Various NSAIDs have been implicated in glomerular disorders leading to proteinuria and/or the nephrotic syndrome, possibly due to some podocyte-specific type of injury [96]. In other instances, interstitial nephritis can occur bacause of NSAID immuno-allergic effects that are most likely unrelated to COX inhibition [97]. Under most circumstances, proteinuria or non-oliguric AKI rapidly disappear upon therapy discontinuation.
Another issue that may impact on the renal adverse effects of NSAIDs is the lack of apparent recognition of renal dysfunction by prescribing physicians. Notably, sudden changes of GFR may go unnoticed if serum creatinine, blood urea nitrogen or serum K+ are not measured during NSAIDs therapy. Thus, the real prevalence of renal untoward effects of NSAIDs may be largely underestimated. A recent systematic review [98] noted a cross-sectional point prevalence of NSAID use of between 8 and 21% in 49,209 patients with CKD, demonstrating that despite guidelines recommending against their use, a substantial proportion of CKD patients continue to receive NSAIDs.
NSAIDs and Arterial Pressure
The pro-hypertensive effects of NSAIDs are believed to stem from three major mechanisms [99]:
-
(1)
Na+ and Cl− retention and increased antidiuretic hormone-mediated water reabsorption at the distal collecting duct
-
(2)
Blockade of the vasodilator effects of PGE2 and PGI2 on the kidney microcirculation
-
(3)
Unbalanced activity of the renin/angiotensin/aldosterone axis, normally regulated by local vascular and tubular eicosanoid biosynthesis.
No effect on blood pressure (BP) has been observed in ASA [100,101,102] and coxib users [103, 104]. Among non-selective NSAIDs, ibuprofen and indomethacin—but not diclofenac—were shown to increase the risk of hypertension in arthritis patients [103]. In a meta-analysis of 19 RCTs including 45,000 patients with arthritis treated for > 4 weeks with COX-2 inhibitors, non-selective NSAIDs or placebo, coxibs caused a weighted mean difference point estimate increase in systolic and diastolic BP compared with placebo and non-selective NSAIDs, and were associated with a non-significantly higher RR of causing hypertension compared with placebo and non-selective NSAIDs [105]. Another meta-analysis of 49 RCTs with 130,000 patients—mostly with arthritis—found that coxibs caused greater hypertension than either non-selective NSAIDs or placebo after at least 4 weeks of treatment. However, the effect was heterogeneous, with a marked BP increase in etoricoxib users and a slight effect in users of celecoxib, valdecoxib and lumiracoxib [106]. The review did not report absolute risk changes or provide numbers needed to treat or harm.
Monitoring the Renal and Nephrovascular Effects of NSAID Therapy
We suggest that patients with cardio-renal risk factors receive a complete nephrological assessment, including calculation of the estimated GFR, age-adjusted renal function, urinalysis, electrolyte and acid–base profiling (acidosis/hyperkalemia), microalbuminuria, proteinuria (if any), concurrent anti-hypertensive therapy (anti-ANG II, anti-aldosterone treatment). Measurement of serum chloride is particularly useful [107]: at < 100 mmol/l, Cl− predicts a setting of metabolic alkalosis (diuretics, hyperaldosteronism); at > 105 mmol/l, it suggests a hyperchloremic metabolic acidosis (renal failure with normal anion gap). Failure of Cl− to increase in the presence of metabolic acidosis with low HCO3− levels implicates an elevated anion gap acidosis, resulting from an unmeasured anion (ketones, lactate, alcohol metabolites, salicylate or other intoxications, sepsis). Both alkalosis and acidosis usually drive significant changes of serum K+, potentially relevant to treatment with NSAIDs, which tend to increase K+ by interfering with prostacyclin-mediated K+ secretion in the distal tubule. If used in conjunction with an angiotensin-converting enzyme inhibitor (ACEi) or ANG II receptor antagonist in a diabetic patient, the risk of hyperkalemia is greatly increased.
Practical Indications
-
(1)
NSAIDs have very modest effects in the healthy kidney and/or younger individuals and are usually well tolerated in subjects with normal renal function.
-
(2)
Any patient with chronic renal disease should be warned against possible sideeffects of NSAIDs, both in terms of renal function and/or blood pressure control. Should a course of NSAIDs be deemed necessary, the following measures should be taken:
-
i.
Obtain a baseline measurement of renal function (i.e. estimated [e]GFR by Cockroft-Gault, CKD-EPI or MDRD equations) and serum K+.
-
ii.
Withdraw any concurrent anti-hypertensive therapy with ACEi or ANG II receptor blockers (known to decrease eGFR in elderly patients with widespread atherosclerotic vascular lesions).
-
iii.
Keep daily doses of the chosen NSAID to the lowest effective level, for no longer than 1 week to 10 days.
-
iv.
Avoid dehydration or concurrent diuretic therapy, unless mandatory.
-
v.
Monitor eGFR and serum K on weekly basis. Virtually all non-selective COX-1 and -2 inhibitors have the potential to induce or aggravate AKI; selective COX-2 inhibitors (rofecoxib, celecoxib) can also affect renal function, whereas NSAIDs with higher COX-2 selectivity (diclofenac, meloxicam) also have renal effects, however not statistically significant.
-
vi.
Closely monitor individuals with increased risk of AKI due to underlying comorbidities (arterial hypertension, diabetes, heart failure, stroke).
-
vii.
Withdrawal of NSAIDs is almost always followed by recovery of renal function, although not all cases of AKI are entirely reversible.
Tables 5 and 6 summarize the indications on the selection of the most adequate NSAID according to the CV, GI and renal risk (low vs. high).
Table 5 Indications on the selection of the most adequate NSAID according to the cardiovascular and gastrointestinal risk, in patients with low renal risk Table 6 Indications on the selection of the most adequate NSAID according to the cardiovascular and gastrointestinal risk, in patients with high renal risk