In this proposal, we outline an Ontario population-wide longitudinal cohort study that will examine the relationship between serum lithium levels and renal outcomes in older lithium users. This study will identify the lithium level above which poor renal outcomes occur, helping clinicians in Ontario and internationally prescribe lithium safely to prevent Chronic Kidney Disease (CKD) and Acute Kidney Injury (AKI). This could allow more patients with geriatric bipolar disorder and depression to benefit from lithium, while also preventing unnecessary psychiatric and medical health service utilization in older mood disorder patients.

Lithium remains the gold-standard medication for the treatment of mood disorders (Yatham et al., 2013), with 30-40% of patients with bipolar disorder responding preferentially to lithium (Geddes et al., 2010; Grof, 2010) and many patients with depression benefitting from it. Late-life bipolar disorder and depression are highly common, affecting up to 0.5-1% and 15% of older adults per year, respectively (Buchtemann, Luppa, Bramesfeld, & Riedel-Heller, 2012; Kessler et al., 2005). An estimated 25% of mood disorder patients are currently aged >65, with that percentage rising rapidly (Sajatovic et al., 2005). Approximately 1% of all Ontarians and 1% of Ontarians aged ≥65 use lithium (Juurlink et al., 2004; Rej, Shulman, et al., 2014) and many more could potentially benefit from this (Geddes et al., 2010; Grof, 2010; Shulman et al., 2003).

However, there have been concerns that lithium may be relatively unsafe in older adults, particularly from a renal (kidney) perspective and especially at higher lithium levels. Chronic Kidney Disease (CKD) and Acute Kidney Injury (AKI) have been particular concerns (Laliberte, Yu, & Rej, 2015; McKnight et al., 2012; Rej, Elie, Mucsi, Looper, & Segal, 2015; Rej, Herrmann, & Shulman, 2012). Our work and others have found that CKD occurs in 42-50% of older lithium patients (Rej, Segal, et al., 2014; Rej, Segal, et al., 2015; Rej, Shulman, et al., 2014; van Melick, Meinders, Hoffman, & Egberts, 2008) and AKI in up to 1.5% of geriatric lithium users (Juurlink et al., 2004; Rej, Shulman, et al., 2014). AKI increases the risk of CKD up to 10-fold (Rej, Shulman, et al., 2014). These renal conditions, particularly CKD, are associated with increased health service utilization (e.g. acute and chronic hemodialysis), with CKD being one of the major contributors to poor quality of life, disability, and mortality (Chin et al., 2014; Murray et al., 2013).

Supratherapeutic serum lithium levels contribute to an increased risk of AKI (Boton, Gaviria, & Batlle, 1987; Close et al., 2014; Rej et al., 2012) and nephrogenic diabetes insipidus (Rej, Segal, et al., 2014; van Melick et al., 2008), both of which can independently increase the risk of CKD up to 3-10 fold (Rej, Elie, et al., 2015; Rej, Segal, et al., 2015; Rej, Senouci, Looper, & Segal, 2013; Rej, Shulman, et al., 2014).

The therapeutic lithium level recommendations for adults have been 0.6-1.2mmol/L, even though levels above 0.8mmol/L are likely toxic to older patients’ kidneys (Rej, Abitbol, Looper, & Segal, 2013; Rej, Looper, & Segal, 2013; Shulman, Mackenzie, & Hardy, 1987). An ongoing ”double disaster” has thus been occurring: there have been relatively high rates of acute lithium toxicity and acute kidney injury (AKI) in 1-1.5% of lithium users/year (Juurlink et al., 2004; Rej, Shulman, et al., 2014) and clinicians have been avoiding prescribing lithium in older adults who may otherwise benefit from it (Shulman et al., 2003; Strejilevich et al., 2011).

Complicating matters, mood disorders themselves are often associated with elevated cardiovascular and chronic kidney disease (CKD) risk. Mood disorders are associated with increased rates of medical comorbidities, including renal disease (Gildengers et al., 2008; Rej, Elie, et al., 2015).Additionally, we have shown that a frequently used alternative for lithium in mood disorders, the atypical antipsychotics, have been associated with acute kidney injury (AKI) (Hwang et al., 2014), as well as CKD (Rej, Shulman, et al., 2014).

In the past year, important studies in the Lancet (Shine, McKnight, Leaver, & Geddes, 2015) and Lancet Psychiatry (Clos, Rauchhaus, Severn, & Donnan, In Press) have tried to examine lithium and renal function in younger adults. However, there remain a number of methodological flaws in the existing literature: many studies have not used a psychotropic-using (or mood disorder) comparison group to reduce concerns about confounding by indication, and did not control for covariates such as diabetes mellitus, hypertension, coronary artery disease, and concurrent medication use. Additionally, older adults have very high pre-existing rates of CKD (Coresh et al., 2007), making them both potentially vulnerable to further decrements in their kidney function (Rej, Abitbol, et al., 2013; Rej, Li, Looper, & Segal, 2014), and having multiple non-lithium related reasons for developing CKD. This makes data from younger samples not necessarily applicable to geriatric lithium users.

There has yet to be a large longitudinal study examining CKD, AKI, and changes in renal function in older lithium users using appropriate comparator groups, controlling for potential confounders, and using laboratory confirmation of renal function (estimated Glomerular Filtration Rate – eGFR).

A potential comparator group could be valproate and atypical antipsychotic users. These are the main alternatives for lithium (Rej et al., In Press) in bipolar disorder (valproate/atypical antipsychotics) and/or as an adjunctive agent in unipolar major depressive disorder (atypical antipsychotics). Using a pharmacologically-defined control group is more reliable than physician mental health diagnostic codes in Ontario’s administrative health databases (Seitz et al., 2012).

Through the linkage of the Ontario Information Laboratory System, a significant innovation that will be available in 2015-2016 at Ontario’s Institute for Clinical and Evaluative Sciences (ICES) is the opportunity to analyze Ontario-wide lithium levels and laboratory measures of renal function. This will allow an even more thorough examination of lithium and renal function using the strengths of ICES data: Ontario-wide population coverage, as well as detailed information about medication use and medical disorders. These innovations will allow for a rigorous examination of the effects of lithium levels on AKI and CKD in older lithium users taking advantage of laboratory confirmation of AKI and CKD, while controlling for important covariates.

If we know more about what the “renal-toxic lithium level” is in older adults, then we could use lithium more safely in more patients that could potentially benefit from it, thereby preventing unnecessary psychiatric and medical health service utilization in many older mood disorder patients in Ontario.