Friday, January 5, 2018

If Your Patient Really Needs Lithium.......

I had the good fortune to work in an intense medical setting for about 22 years where I was responsible for the medical care of my patients.  In that setting I had access to excellent specialists, in fact some of the finest physicians I have seen anywhere.  They were typically involved when I had identified a medical problem that potentially complicated the psychiatric care of my patients.  In those cases - no matter what the problem was it usually came down to the question: "If your patient really needs this medication we need to continue it."  One of the most significant complications was renal failure of varying degrees while taking lithium.  Some patients on my unit were hospitalized because they had been stable on lithium but were developing renal failure and needed to be tried on another agent.  In some cases they experienced severe associated complications including delirium either from the bipolar disorder or medications used in the transition.  In some cases, they needed dialysis and renal medicine consultants would advise me on the lithium dose to try while they were receiving hemodialysis.  All of this experience led me to have a very low threshold for recognizing and acting on potential adverse effects of lithium as early as possible. 

At various points in my career, there was a question of renal failure occurred in cohorts on lithium or it was due to a different cause.  As an example, the Lithium Encyclopedia from 1983 (1) stated concisely: "There have been a growing number of reports of morphological and functional kidney damage associated with lithium use.  The actual incidence of lithium nephrotoxicity is not known.  Most researchers believe that irreversible damage is not widespread but that risk increases with length of treatment and serum lithium level."

At the time there were opinions that up to 20% of patients on long term lithium maintenance had morphological changes and functional changes primarily with water absorption but no reduced glomerular filtration rate (GFR) or renal insufficiency.  My favorite renal and electrolyte disorders text cited lithium as a compound that caused vasopressin resistant nephrogenic diabetes insipidus (NDI) (2).  More recent evidence suggests that 15-20% of patients on lithium develop a progressive decrease in GFR typically does not progress to end stage renal disease (ESRD) and dialysis but in some cases it clearly does.  These generalizations are usually based on low numbers of patients who have been identified as having a specific creatinine or GFR.  A creatinine of 2.5 mg/dl predicted progression to ESRD in most patients.  A larger study in Sweden of 3369 patients, suggested a lower threshold. In that study,  13 patients had been off lithium for 2 years before starting dialysis and 11 had a creatinine of > 1.4.  Some were lower but there had been a progressive increase from baseline suggesting the rate of change is important.  Additional factors such as the use of NSAIDS and ACE inhibitors like lisinopril as well as other intrinsic kidney diseases can be a factor.

There have been a recent increase in studies that look at the side effects of lithium and in one case (6) compare lithium side effects to other typical mood stabilizers (valproate, olanzapine, questiapine).  One of the studies uses descriptive statistics and the other two involve calculations of hazard ratios for specific risks.  The results are fairly uniform in their conclusions and have been known in most psychiatry training programs for the past 10-20 years.  The articles here use the Kidney Disease Outcomes Quality Initiative (KDOQI) staging of chronic kidney disease published in 2002.   The range is from stage 1 (eGFR ≥ 90 ml/min/1.73 m2 ) to stage 5 (eGFR < 15 ml/min/1.73 m2 or dialysis).  For the purpose of these papers Stage 2 is an eGFR of 60-89 and Stage 3 is 30-59. As previously noted progression can occur in some cases even years after the lithium has been stopped.

The most interesting of the three papers is probably from Hayes, et al because of their strategy to look at lithium toxicity but in the context of other commonly used mood stabilizers (valproate, olanzapine, quetiapine) and the associated toxicities of these other mood stabilizers.  As a result they have a table entitled Table 2. Adverse effects during maintenance treatment that looks at these mood stabilizers and CKD (stage 3 and 4), hypothyroidism, hyperthyroidism, hypercalcemia, Type 2 diabetes mellitus, cardiovascular disease, hypertension, hepatotoxicity, and weight gain (greater than 7% and greater than 15%).  Hazard ratios care calculated for all of these adverse effects using lithium as the reference.  The number of events in each category is relatively low compared with the total events in a large clinical sample.  The adverse events described were expected.  They estimated the rate of severe CKD (stage 4 or 5) as 1 in 100 person years at risk.  The estimated rates of stage 3 and stage 4 CKD were 9 in 100 per years at risk.  Hepatotoxicity was rare in the sample and was elevated in the quetiapine group.  Weight gain was most significant for olanzapine, quetiapine, and valproate relative to lithium.  Olanzapine also had the highest rate of new onset hypertension.

None of this information deters me from offering lithium to people who I think are good candidates.  The rationale is that lithium can be a life changing medication, especially for people who have generally never achieved mood stabilization with all of the typical medications prescribed for mood disorders these days.  Informed consent should include this potential complication.  I generally advise people that there is a 40% risk of NDI of varying severity, the need for ongoing testing, the need to avoid additional medications that may complicate the use of lithium, and finally close self monitoring of both side effects and any situation that can complicate lithium therapy like conditions leading to dehydration.  I will add the risk of mild CKD and explain to the patient what that means.

One of the interesting aspects of these articles is that ongoing screening for patients on lithium therapy varies considerably from setting to setting.  National guidelines seem to help.  In these large scale studies there is not enough discussion of work at the clinical level incorporating the red flags from the research.  For example it is obviously important to know about new medications, especially ACEIs, NSAIDs, and serotonergic medications.  Medications that may complicate interpretation of lab finding like Vitamin D and Calcium supplementation in the case of hypercalcemia are also important.  The review of systems is also important anywhere along the spectrum of side effects including neurological, cognitive, gastrointestinal, and renal symptoms.  I generally advise patients that increasing nocturia, polyuria, and polydipsia with associated laboratory finding increases the risk of irreversible renal changes including a low risk of progression to ESRD.  The presence of edema may indicate the rare onset of a nephrotic syndrome that can occur is some people with acute treatment.

The good news is that even now there is an accumulating literature on the potential complications of lithium therapy and some general ideas about what psychiatrists should do about it.  Although the methodology of these studies varies significantly they generally come to similar conclusions about the effect of lithium on renal function, thyroid function, and calcium metabolism.  All of these systems require close monitoring by assessing the patients clinical status and lab testing.  Numerous changes in clinical status should result in departures from any monitoring routine.  The critical elements is communication with the patient and education about when the psychiatrist should be contacted.

George Dawson, MD, DFAPA 


1:  Jefferson JW, Greist JH, Ackerman DL.  Lithium Encyclopedia for Clinical Practice.  American Psychiatric Press, Inc. 1983: p151.  

2:  Schrier RW (ed).  Renal and Electrolyte Disorders, 2nd Ed.  Little, Brown and Company, Inc.  Boston 1980: p. 31.

3:  Lerma VE.  Renal toxicity of lithium. In UpToDate.  Sterns RH, Sheridan AM (eds) (Accessed on January 2, 2018).

4: Bendz H, Schön S, Attman PO, Aurell M. Renal failure occurs in chronic lithiumtreatment but is uncommon. Kidney Int. 2010 Feb;77(3):219-24. doi: 10.1038/ki.2009.433. Epub 2009 Nov 25. PubMed PMID: 19940841. (see open access text).

5:  Dineen R,  Bogdanet D,  Thompson D, Thompson CJ,  Behan LA,  McKay AP, Boran G,  Wall C, Gibney J, O’Keane VO, Sherlock M.   Endocrinopathies and renal outcomes in lithium therapy: impact of lithium toxicity.  QJM: An International Journal of Medicine, Volume 110, Issue 12, 1 December 2017, Pages 821–827,

6: Hayes JF, Marston L, Walters K, Geddes JR, King M, Osborn DP. Adverse Renal, Endocrine, Hepatic, and Metabolic Events during Maintenance Mood Stabilizer Treatment for Bipolar Disorder: A Population-Based Cohort Study. PLoS Med. 2016 Aug 2;13(8):e1002058. doi: 10.1371/journal.pmed.1002058. eCollection 2016 Aug. PubMed PMID: 27483368.

7: Shine B, McKnight RF, Leaver L, Geddes JR. Long-term effects of lithium on renal, thyroid, and parathyroid function: a retrospective analysis of laboratory data. Lancet. 2015 Aug 1;386(9992):461-8. doi: 10.1016/S0140-6736(14)61842-0. Epub 2015 May 20. PubMed PMID: 26003379.


  1. The biggest mistake psychiatrists, and other MDs who more often recklessly prescribe Lithium without being fully educated to the Risk/Benefit profile, is altering dosages simply per lithium levels.

    We treat people, not numbers. Be candid Dr D, how many of our colleagues reflexively change scripts simply because the level is "too low". I had a patient put into delirium because some nitwit nephrologist who was treating my patient's nephrolithiasis double her dose without consulting me first because her therapeutic level of 0.3 for TWO YEARS without any features of relapse was "low" to him.

    Hey, I don't write for oncology meds, why the hell are non psychiatrists writing for Lithium?!

    Joel Hassman, MD

    1. Agree completely that there is too much attention paid to arbitrarily defined "therapeutic levels". Best example I can think of is different therapeutic changes for different labs. I really doubt that any of these levels were based on therapeutic patient outcomes. I have also treated patients in remission with what were considered "subtherapeutic levels" - generally less than 0.5. My speculation is that if the low levels work there are fewer side effects and less chance of nephrotoxicity.

      Have not had anyone change the dose on me but had one incident of "toxicity" when another MD decided to check the level after the AM dose. Had another case of a utilization reviewer trying to embarrass me by a hostile call to nursing staff: "Tell Dr. Dawson to get that patient a STAT appointment for a clinic where they can adjust the lithium levels." The real implication was that I did not know what I was doing with the patient or lithium. He failed to see the labs on the day of his call and a lithium level of 0.7.