NSAIDs, Blood Pressure and the Kidneys: Who Should Be Careful

In the kidneys prostaglandins act as a regulator. They dilate the small vessels that bring blood to the glomeruli, promote the excretion of sodium and water, and modulate the renin-angiotensin system. When NSAIDs inhibit their production, the renal vessels constrict and the kidney retains more sodium and water. Circulating blood volume and vascular resistance rise, which translates into higher blood pressure. In healthy people with a well-filled vascular bed, the kidneys usually compensate for this change. The problem appears when renal blood supply is already prostaglandin-dependent, that is in dehydration, heart failure, liver cirrhosis, chronic kidney disease and in older people.

The same mechanism explains why NSAIDs weaken many blood pressure drugs. ACE inhibitors, sartans, diuretics and beta-blockers work partly through effects on sodium-water balance and the renin-angiotensin system, in which prostaglandins act as a mediator. When they are blocked, the effectiveness of antihypertensive treatment may fall, so a person who 'keeps blood pressure in check' may see higher values after a few days of an NSAID. We describe this as a mechanistic explanation: it is supported by the clinical observations described below, but the mechanisms themselves can be more complex.

How should this be read? First, 5 mmHg is an average from very different trials: the confidence interval is wide (1.2-8.7), so in some people the effect will be close to zero and in others much greater. Second, 5 mmHg for one person is not much, but across a population a shift of that size matters for the risk of stroke and heart attack, as the authors themselves stress in their wording about possibly increased morbidity. Third, this is an analysis from 1994, in which older drugs and other regimens dominated; it shows the direction of the effect but is not the last word.

Also notable is the observation that the effect of NSAIDs was greatest in people taking beta-blockers. This is a conclusion from pooled data, so it must not be generalised to every patient, but in practice it means that anyone treating hypertension should regard NSAIDs as a drug that can affect the effectiveness of that therapy.

This is an important addition, because it shows the effect on blood pressure in a modern, well-designed trial with round-the-clock measurement, not only a single reading in the clinic. The context must be remembered, though: these were prescription doses, higher than the typical over-the-counter doses, and the participants were middle-aged and older people with joint disease and increased cardiovascular risk. The result does not mean that one ibuprofen tablet from the pharmacy will raise blood pressure by 3.7 mmHg. It means that with regular, months-long use of prescription-dose ibuprofen, blood pressure rises on average and hypertension develops more often than with celecoxib.

Differences between drugs are therefore sometimes smaller than the differences between short and long use and between low and high doses. In the review we describe in the article on ibuprofen and alcohol, the authors point out that NSAID risk is continuous and depends on total exposure, that is on dose and duration of use.

First of all, NSAIDs can cause acute kidney injury, that is a sudden worsening of kidney function, usually in situations where renal blood supply depends on prostaglandins: dehydration, heat, vomiting and diarrhoea, heart failure or after heavy exertion. In most cases the injury is reversible after stopping the drug, but it has to be recognised. The second question is chronic kidney disease (CKD): whether long-term NSAID use speeds its development or progression.

How should this be interpreted? The meta-analysis covers only observational and interventional studies with very high heterogeneity (I2 above 90%), and people chronically taking NSAIDs differ from others: they more often have joint disease, diabetes and hypertension, so part of the association may come from underlying diseases and other drugs. A cautious conclusion is that there is a consistent signal linking chronic NSAID use with CKD, strongest in people who already have kidney disease. Causality is not proven. We describe creatinine and eGFR results in our entry on creatinine and eGFR.

The term 'triple whammy' describes the simultaneous use of a diuretic, a renin-angiotensin-aldosterone system inhibitor (an ACE inhibitor or sartan) and an NSAID. Each of these elements disturbs the blood supply to the glomeruli in a different way: the diuretic reduces circulating blood volume, the ACE inhibitor or sartan dilates the vessel leaving the glomerulus, and the NSAID constricts the vessel entering it. Together they can lead to a fall in glomerular filtration, especially when dehydration is added.

Two independent sources point in the same direction, but with important caveats. The absolute risk was low (an incidence of acute kidney injury of 7 per 10,000 person-years in the cohort), double combinations showed no rise, the meta-analysis rests on case-control studies with low certainty of evidence and high heterogeneity, and a rise in mortality was not shown. The most practical conclusion concerns the first weeks: the greatest risk occurs at the start of concurrent use, so that is when to be most watchful for dehydration and symptoms.

The first thought is paracetamol, because it does not affect the kidneys or stomach the way NSAIDs do. It turns out, however, that it too is not entirely neutral for blood pressure, as shown by a randomised trial in a separate population of people with hypertension.

This trial concerns a dose of 4 g a day for two weeks, that is the maximum dose, not a single tablet, and the participants were people with diagnosed hypertension. It does not mean that paracetamol is 'worse' than NSAIDs, but that no oral painkiller is entirely indifferent to blood pressure if used regularly and at a high dose. We write about paracetamol dosing, the liver and alcohol in the article on the maximum paracetamol dose.

The practical conclusion: for local pain in osteoarthritis, NSAIDs applied to the skin have evidence of effectiveness, albeit moderate and for a minority of people, and in the studies no clear difference in systemic adverse events from carrier was observed. The review has limitations, however: the quality of evidence on systemic safety is very low, much data from completed studies remained unpublished, and for other pain conditions (for example acute sprains, muscle pain) the chronic-pain review cannot be applied. People with CKD or in risk groups should discuss gels with a doctor or pharmacist too.

NSAIDs remain effective drugs, and for a healthy person short, occasional use is usually justified. The data show, however, that with hypertension, kidney disease, in older age and when combined with diuretics and ACE inhibitors or sartans, it is worth changing the default strategy. This is not about panicky avoidance, but about three decisions: whether the pain really needs a systemic drug, whether a local drug or a non-drug method can be used, and whether someone who knows the full medicine list has confirmed that this combination is safe for us.