Lithium continues to be a first line treatment for mood conditions such as bipolar disorder, and rightfully so, as it continues to be unmatched in its ability to stabilise moods (see, for example, this recent blog about reduced hospitalisations being associated with Lithium use).

However, its negative impact on kidneys is well documented and, if you belong to any serious mental illness support group, you will know a person or three whose kidneys have been impacted by lithium. And so, as patients, we often ask ourselves “kidneys or mental health?” And this is not an arbitrary question to those of us who live with a serious mental illness. We are not the chicken whose contribution to breakfast is an egg; we are the pig whose contribution to breakfast is a sausage. We have serious skin in the game!

Inching us forward in this quest for answers, van der Aa et al. (2026) sought to quantify how lithium affects kidney function over ten years.

Methods

This study was based in the Netherlands, and the criteria for inclusion were:

• patients living with a mood disorder,
• have been prescribed lithium,
• were active in the medical system, and
• had a follow-up of more than a year.

They ended up having 196 patients of which 42% were male, with a mean age of 51.1 years at start. The measurements to be used were based on estimated glomerular filtration rate (eGFR: a key indicator of how well the kidneys are filtering waste from the blood) and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.

The investigators tracked:

• the length of time that a patient was on lithium therapy,
• ongoing blood serum concentrations,
• whether the patient was on any other medications (comedication), and
• whether any patients stopped lithium during the ten year period.

Patients who stopped lithium during the study were monitored for any further decline in eGFR.

The researchers had their work cut out for them. Patients who stopped using lithium were compared against those patients who continued lithium treatment. And, patients who stopped lithium treatment due to eGFR decline were compared against patients who ceased lithium treatment for other reasons. Medications known to influence eGFR decline except lithium were carefully excluded. The research team then tracked how lithium was used, the dose and the effect of discontinuation on kidney function.

Results

The median follow up was 8.8 years and the investigators found that there was a significant decline (p < 0.001) in eGFR at inclusion versus the eGFR at follow up. The trajectories of eGFR in most patients were linear however eGFR trajectories were not linear in patients who were either pregnant or admitted in clinical hospitals. And so patients’ results while hospitalised and when pregnant were excluded. The median decline in eGFR in patients was 0.79 mL/min/1.73 m2 per year; 11% of the participants had > 2.5 mL/min/1.73 m2 eGFR decline per year.

The results showed a relationship between lithium concentration and eGFR decline when controlled for age, sex and duration of lithium therapy at baseline. Furthermore, there was a significant association between median duration (which was 16 years) and eGFR reduction. Other medications known to increase the risk of kidney decline were not found to have a significant association with eGFR decline when the results were corrected for sex and age.

Thirty nine of the 196 patients stopped lithium treatment. In 18 instances the reasons for discontinuation were due to documented side effects including kidney decline, nephrotoxicity (kidney damage), complaints of polyuria (excessive urination connected to lithium usage) or psychiatric remission. In three other instances the reasons for discontinuation were not documented. The paper does not seem to account for the remaining 18 discontinuations.

The decline of eGFR observed in long term lithium users was more than the decline observed in the general Netherlands population, suggesting that the increased decline was linked to the long term use of lithium. The research team also noted that before ceasing lithium, the patients who stopped using lithium experienced similar decline in eGFR to those who stayed on lithium during the tenure of the study. Of the 39 patients who stopped lithium, only 33 had follow up information which was used to determine this finding. Although the decline slowed (from 1.3 to 0.27 mL/min/year) in those who ceased lithium intake, it was not significant for the purposes of this study when tracked during this ten year period. Another possible reason for them concluding that there was no significant slowing in kidney decline is that they had a small number of participants in their studies and then an even smaller number of participants stopping lithium. It could also be possible that they did not see statistically significant improvement because they could not track the long term differences.

Moreover the research team found that there was a significant relationship between higher serum concentrations and eGFR decline, meaning there is an association between lithium treatment and kidney function decline. The team found no evidence that comedication between lithium and nephrotoxic medications could be worsening the decline of kidney function, but the jury is still out on this one.

Conclusions

The authors concluded:

This study provides further evidence that eGFR decline occurs in most long-term lithium users. Of this cohort, 59% of the participants had faster eGFR decline than the mean decline in the general population. Lithium exposure, quantified as mean serum lithium concentration, could be a contributing factor in this decline.

Strengths and limitations

I agree with the research team that their study was important because they had long follow up time and their cohort had been on lithium for a long period, they could use the results from bi-annual monitoring of patients who are on lithium to construct a trajectory of their kidneys.

The study also had a number of limitations starting with the fact that, as an observational study, it is not possible to draw conclusions about a causal relationship between lithium and kidney function. There are further limitations of the study that I found to be important. The first is that the researchers could not verify whether patients had waited 12 hours after lithium intake before taking lithium measurements and this could have influenced the results of the serum lithium concentration captured. The second is that the researchers predominantly relied on medical reports after the baseline to determine comedication. We just hope there was no attrition of data as it was transferred from patient to the doctor and then doctor to the researcher. I suspect there could have been just because consultations are so short that there is no time to relay everything to the doctor. So you leave with the same prescription for medications that fill your medicine cabinet, taken on an “as and when I see necessary basis”.

Other studies have hypothesised that the presence of comorbidities (hypertension, diabetes etc) could be a possible reason for the eGFR decline (Clos et al, 2015), however this study did not have comorbidities as a variable so we cannot draw any conclusions on that possibility. We hope to see researchers including these disease variables in future studies. van der Aa et al. also had a limited number of nephrotoxic medications to be able to conclusively determine whether they could be the actual cause of the eGFR decline. The research team recognises that one of the study limitations, like other studies investigating lithium’s impact on the kidneys, was it did not have a large enough cohort. This makes it harder to have confidence in its findings, which is concerning given the seriousness of these side effects and the importance of lithium as a treatment.

Implications for practice

I found the study impressive in that it was robust in its design and therefore meaningfully contributed to the knowledge around the impact of lithium use on the kidneys. It also followed the patients instead of collecting data and then working backwards, which allowed them to also see what happens to patients who stay on lithium versus those who stop lithium treatment.

This study highlights that we do not yet know whether stopping lithium once kidney function is significantly reduced is helpful. And what of the mood instability that could occur as a result of stopping lithium treatment?

Although probably seen in practice, these investigators add to the body of knowledge supporting that clinicians should try to pursue the lowest lithium concentration necessary to achieve mood stability and general optimal clinical outcomes. van der Aa et al. support the monitoring of lithium concentrations as per the current guidelines and measuring kidney function in a manner that is individualised. Importantly, they make no recommendation on the course of action to be taken when people who are taking lithium begin experiencing kidney decline.

The research team has many recommendations for future studies including prospective designs that seek to capture the biomarkers that can work as surrogate measures.  These surrogate measures can be used to capture future kidney function decline thereby removing the need for expansive decade long clinical studies. These can also be used to confirm whether there is a point of no return when it comes to kidney decline.

Now remember, there are usually two people in the psychiatric room during consultation. One (the practitioner) should take care to use best practice when prescribing the dose of lithium to be taken by the patient. The other (the patient) should remember that they are the proverbial pig and not the chicken, so it will do us a whole lot of good to be our own best advocate in that consulting room.

Statement of interests

The author of this blog (Nomhle Nhlapho) declares that she stopped lithium within 3-6 months of starting lithium treatment due to side effects from the medication.

Edited by

Edited by Simon Bradstreet.

Links

Primary paper

van der Aa, M. J., Zittema, D., Doornebal, J., Hartong, E. G. T. M., Bisseling, E. M., Dammers, J., Klumpers, U. M. H., Kerckhoffs, A. P. M., Kupka, R. W., & Nijenhuis, T. (2026). A Significant Decline of Glomerular Filtration Rate in the Majority of Long-Term Lithium Users: Results of a Dutch Prospective 10-Year Cohort Study. Bipolar Disorders, 28(2), e70082.

Other references

Clos, P. Rauchhaus, A. Severn, L. Cochrane, and P. T. Donnan. (2015). Long-Term Effect of Lithium Maintenance Therapy on Estimated Glomerular Filtration Rate in Patients With Affective Disorders: A Population-Based Cohort Study, Lancet Psychiatry 2, no. 12: 1075–1083

Photo credits

• Photo by Olga DeLawrence on Unsplash
• Image by Europeana on Unsplash
• Photo by Aakash Dhage on Unsplash
• Photo by Roberto Sorin on Unsplash