Lithium and a Falling eGFR: Is There a Number at Which You Must Stop, and Why Not Wait for 30 to Decide?

The letter from primary care is short. The eGFR is 46, the patient now has “CKD stage 3,” and should lithium be stopped? Two reflex answers follow. One is to stop lithium now, because the kidney number has crossed a line. The other is to carry on until the eGFR reaches 30, because that is the number in the label. Both treat the decision as a threshold. It is not one. No number above the label’s boundary forces a stop, and waiting for that boundary means making the most important decision in long-term lithium care at the moment the evidence is least settled.

Is there an eGFR at which lithium must stop?

No. The label draws one boundary, at a creatinine clearance below 30 mL/min. Every other number you will hear quoted is a guide for when to act, not a rule that decides for you.

The US lithium label does not recommend lithium in severe kidney impairment, which it defines as a creatinine clearance below 30 mL/min. It asks for kidney function to be checked before and during treatment, and for lower doses, slow titration and frequent levels when kidney function is reduced. It does not give an eGFR at which lithium has to stop. Here is what the commonly quoted numbers actually mean.

If not a number, what should make me act?

A sustained fall across several results. Read eGFR as a series, and repeat a single low result once any reversible cause has passed.

Lithium’s effect on the kidney is slow, and slow harms are the ones clinical systems handle worst. Each result is seen by a different person, each is a little lower than the last, and each is still “acceptable.” A psychiatry editorial on eGFR in lithium monitoring suggested investigating a progressive fall across three or more serial tests. The 2026 expert review makes the same point: decisions should follow the trajectory, not a single value.

Two traps sit on either side of the trend.

1. One low result can be temporary.A fall after a toxicity episode, dehydration or a new interacting drug may partly recover. In a Swedish series of lithium-treated patients with at least one eGFR below 30, about half had a transient fall followed by improvement. Repeat the result once the cause has passed.
2. A slow fall can hide in the normal range.eGFR falls with age in everyone. The question is whether this patient’s rate is steeper than age alone explains, and that is exactly what a nephrologist can help judge.

For when to involve nephrology, the UK NICE kidney guideline gives criteria that apply to any adult with chronic kidney disease, and lists lithium among the medicines that need eGFR monitoring at least once a year.

Who is most at risk?

Patients with long exposure, older age, higher levels, previous toxicity episodes and other kidney risks. Most of these you can see coming.

Moderate decline is common. A 2026 Dutch prospective cohort of 196 long-term users found that most lost kidney function faster than the general population, and that higher lithium levels and longer treatment went with faster decline. Severe disease is the tail of the distribution, not the body: a Swedish laboratory study with follow-up of more than 35 years found severe chronic kidney disease uncommon overall, with most cases after 70.

Does stopping lithium actually help the kidney?

The best direct evidence says it can slow the decline. How much it helps once kidney function is badly reduced is still disputed.

Read closely, the Dutch and Swedish findings are less opposed than they first appear: the Dutch study had few patients who stopped, and its numbers moved in the same direction. The real disagreement is about patients whose kidney function is already badly reduced. That is exactly why the decision should not be left until then.

So when, and how, should I decide?

This month, once a sustained fall is confirmed, with the patient and a nephrologist, and in writing.

1. Plot every eGFR you can find, with dates.Decide on the trend, not on the latest number.
2. Rule out a reversible cause.A toxicity episode, dehydration, an NSAID, or a new ACE inhibitor, ARB or diuretic.
3. Check the rest of the picture.A urine protein test, blood pressure, glucose, calcium and a 12-hour lithium trough.
4. Review whether the current level is needed.Lower is kinder to the kidney. Use the lowest level that keeps the patient well.
5. If the fall is sustained, involve nephrology.Send the eGFR series with dates, years on lithium, the usual trough, any toxicity episodes, and how well lithium has worked.
6. Make the continue-or-stop decision with the patient, in writing.Record the reasons, the patient’s view, and what would change the decision.

Is there a catch?

Three. The evidence is still split at low eGFR, stopping is itself a risky intervention, and heavy proteinuria is a different problem.

1. The evidence at low eGFR is not settled.One study found the biggest benefit of stopping in patients with the lowest kidney function; reviews expect the least. Nobody should tell a patient with an eGFR in the 30s that stopping is pointless, or that it is certain to help.
2. Stopping lithium is a psychiatric intervention.The Swedish authors themselves stress that kidney benefit must be weighed against relapse and suicide risk, and call for a patient-centred, not a kidney-centred, decision. If lithium is stopped, taper rather than stop abruptly, agree the replacement plan first, increase psychiatric review, and keep monitoring kidney function, because decline may continue for a time.
3. Heavy proteinuria is not the typical lithium picture.The usual lesion is tubulointerstitial, and the urine is often unremarkable. The label does report nephrotic syndrome on lithium, with remission after stopping. New heavy proteinuria or oedema goes to nephrology promptly, with lithium named in the referral.