Soham Rej MD, MScGeriatric Psychiatry Research Fellow, University of Toronto

Geri-PARTy Research Group, Jewish General Hospital, McGill University, Montreal, Canada

CAGP 2016

Faculty/Presenter Disclosure Faculty: Soham Rej

Research Funding:

Canadian Institutes of Health Research (CIHR), Fondsde Recherche Quebec Santé (FRQS), Ontario Mental Health Foundation, McGill University, Charitable Donations to the Jewish General Hospital Division of Geriatric Psychiatry

Relationships with commercial interests:

None

Disclosure of Commercial SupportThis program has received no in-kind support from outside organizations.

Potential for conflict(s) of interest:

None

Mitigating Potential BiasNo Commercial Bias

Objectives Understand the clinical challenges of medication

tolerability and cognitive dysfunction in late-life bipolar disorder;

To appreciate new Canadian and International data about these aspects of late-life bipolar disorder

Engage in a discussion on how to interpret and apply these findings (as far as is possible) in clinical practice.

2016 Update in Late-Life Bipolar Disorder! Exciting year

Post-hoc geriatric analysis of two RCTs found lurasidone to be superior to placebo (n>140)

Large geriatric RCT for lithium and valproate compared both agents against each other) for acute mania and hypomania – both effective (n>200)

Geriatric-specific Bipolar guidelines to be included in CANMAT 2017 – more details next year!

Sajatovic et al. 2016 J Clin Psych; Young et al. in press

Cognitive Dysfunction in Late-Life BD >30% of late-life BD pts have cognitive dysfunction

Visio-spatial/executive, not AD, not necc. progressive over 2-5yr f/u, Cholinesterase Inhibitors not found to be helpful

Cognition associated with low psychosocial functioning

Anticonvulsants have high cognitive risk vs. lithium

OR 1.25-2 of dementia - valproate most associated, carbamazapine , antipsychotics also implicated

?less white matter integrity

Aside from Lithium, Lamotrigine may be somewhat protective

May be driven by lamotrigine’s effect in bipolar depression

Shulman et al. 2005 – J Clin Psych; Kessing et al. 2008 – Archives Gen Psych; Dols et al. 2013 – Int Clin Psychopharm;

Tsai et al. 2007; Daban et al. 2006; Gildengers et al. 2014 – Bipolar Disorders; Gerhard et al. 2015 – Br J Psychiatry

Khan et al. 2004 – J Clin Psych

Statins and cognition in BD? Statins have protective effects in cognition, white

matter integrity, in general

Theoretically statins have anti-inflammatory, pro-endothelial, anti-oxidant effects

Medical comorbidity (for which statin is used) associated with worse cognition

However, controversial ?negative effects on cognition and mood, interferes with lithium’s peripheral effects

Are Statins associated with cognition in BD?

Answers from COG-BD study of 143 BD patients…

Nadkarni et al. 2015 Alzheimers Dement.,

Main ResultsCognitive Domain Statin Users

(n=48)

Mean (+/-SD)

Non-Users

(n=95)

Mean (+/-SD)

Statistics

Language (n=142) -0.56 (+/-0.84) -0.32 (+/-0.82) t(140)= 1.67, p=0.10

Memory (n=137) -0.81 (+/-0.90) -0.59 (+/-0.90) t(135)= 1.36, p=0.18

Executive Function

(n=138)

-0.85 (+/-1.03) -0.57 (+/-1.03) t(136)= 1.51, p=0.13

Visuospatial (n=139) -0.96 (+/-1.24) -0.62 (+/-0.98) t(137)= 1.74, p=0.084

Global (Composite of

all domains tested)

(n=127)

-0.60 (+/-0.69) -0.49 (+/-0.68) t(127)= 0.80, p=0.42

Cognitive Domain Z-Scores and Statin Use in

Older Bipolar Disorder Patients– Univariate

Analysis (n=143)

Tendency of lower cognition in statin users (non-significant) – No association

after controlling for cardiovascular covariates

Conclusions Statins not associated with cognitive dysfunction (or

cognitive benefits) in BD

?Positive and negative theoretical effects cancel out

Any negative “association” likely unrelated to statins

Confounding by indication: statin users more medically ill

Based on this data, statins likely safe to use in BD patients

Medication Tolerability

Lithium: gold-standard treatment for Bipolar Disorder (BD) Useful in Treatment-Resistant Depression,

Anti-suicide effects, Emerging potential role for dementia,

several neurological conditions

30 to 40 % of BD patients respond better to lithium compared to other medications

Many patients stable on lithium are reaching their 70s/80-90’s. >50% of BD patients aged>60 by 2030

Geddes et al. 2010 – Lancet; Al Jurdi et al. 2008 – AJGP; Grof et al. 2002

Lithium and the Kidney

Despite this, trends away from lithium and towards prescribing antipsychotics and antiepileptics in parts of the world, especially North America Only 8-15% of late-life BD pts use Li. <1.5% of

pts on Li monotherapy. Fears of renal effects (esp. Chronic Kidney

Disease), a common reason to avoid/stop lithium.

CKD associated with morbidity and mortality.

Recent high-quality studies in younger/mixed-aged adults have been mixed r.e. is lithium associated with kidney disease?

Geriatric data relatively low quality sample sizes <50-100, cross-sectional

Rej et al. In Press – Int J Geri Psych; Kessing et al 2015 – JAMA Psych, Shine et al. 2015 – Lancet;

Lithium and the Kidney

Methods So we conducted a Nested case-control study using

province-wide data from Ontario, Canada

Patients aged ≥66 with mental health visit between 2003-2008 followed until up to 2014.

Compared 21,741 incident CKD cases and 86,930 age-and sex- matched non-CKD controls for exposures in 5-year lookback: lithium vs. valproate vs. neither.

Conditional logistic regression to assess Incident CKD risk in exposure groups, controlling for 13 covariates, compared to no lithium/no valproate exposure (reference group).

Results

Lithium use associated with increased risk of incident CKD relative to non-use (adjusted OR 1.76 [1.41-2.19])

Valproate use was not (adjusted OR 1.03 [0.81-1.29]).

Parameter

Incident CKD

*Unadjusted

Incident CKD

**Adjusted

OR [95%CI] OR[95%CI]

Lithium only (n=529) 1.37 [1.12-1.67] 1.76 [1.41-2.19]

Valproate only (n=498) 1.21 [0.98-1.49] 1.03 [0.81-1.29]

Lithium and Valproate (n=91)*** 1.98 [1.28-3.06] 1.95 [1.19-3.18]

Neither Lithium nor Valproate (n=107,553,

reference)1.00 1.00

Results: Other Risk Factors All associated with a marked elevation in CKD risk

(ORs 1.37 to 6.4)

Diabetes mellitus,

Hypertension,

Ischemic heart disease,

nephrogenic diabetes insipidus,

acute kidney injury,

Medications: loop diuretics, hydrochlorothiazide, Angiotensin Converting Enzyme Inhibitors and Angiotensin Receptor Blockers

Atypical antipsychotics: OR=1.16

Conclusions Taken together, it seems that in this community-based

sample, lithium is associated with an almost 2-fold increase in CKD risk in older adults.

Contrasts with smaller studies in academic centres Reason for discrepancy needs to be examined further:

Safer monitoring + prescribing in academic centres?

In the meantime… Lithium superior in many older mood disorder patients and

could continue to be used cautiously where the benefits outweigh the risks.

Ensure adequate monitoring of lithium levels and renal function when lithium is used. Manage cardiovascular risk factors

Next Steps Larger population-based studies, using laboratory

confirmation of renal function (eGFR) and lithium levels to assess whether prescribing and monitoring practices and explain poor renal outcomes

Our lab is also looking into pharmacological and biomarker approaches to prevent and treat renal disease in older lithium users

Thank You!

Extra Slides

Baseline Characteristics StatinCharacteristic Statin Users (n=48)

Mean (SD) or % (n)

Non-Users (n=95)

Mean (SD) or % (n)

Statistics

Age (yrs) 68.2 (+/-9.6) 62.4 (+/-8.6) t(141)=3.70, p<0.001

Female Sex 45.8% (n=22) 74.7% (n=71) χ2(1)=11.7, p=0.001

Education (yrs) 15.0 (+/-2.82) 15.2 (+/-2.78) t(141)=0.36, p=0.72

Caucasian Race 93.8% (n=45) 89.5% (n=85) Fisher’s Exact p=0.40

Study Site (Toronto) 39.6% (n=19) 30.5% (n=29) χ2(1)=1.17, p=0.28

Study Site (Pittsburgh) 60.4% (n=29) 69.5% (n=66)

BD Illness and Medical Factors

BD Type 1 72.9% (n=35) 81.1% (n=77) χ2(1)=1.24, p=0.26

BD Type 2 27.1% (n=13) 18.9% (n=18)

Duration of BD Illness (yrs) 38.1 (+/-14.2) 35.7 (+/-12.5) t(141)=1.4, p=0.30

# of Mood Episodes 25.5 (+/-35.2) 22.3 (+/-29.3) t(141)=0.57, p=0.57

Alcohol Dependence 6.3% (n=3) 23.2% (n=22) χ2(1)=6.32, p=0.012

YMRS 2.31 (+/-2.50) 2.43 (+/-2.52) t(141)=0.27, p=0.79

HRSD 4.31 (+/-2.87) 3.78 (+/-2.71) t(141)=1.09, p=0.28

CIRS-G Total Score 9.96 (+/-4.04) 7.34 (+/-3.48) t(141)=4.02, p<0.001

CIRS-G Vascular Sub-Score 2.02 (+/-0.64) 0.89 (+/-0.95) t(141)=8.41, p<0.001

FRSP (n=121) 12.4 (+/-5.33) 8.49 (+/-4.44) t(119)=4.29, p<0.001

Medication Use

Lithium 22.9% (n=11) 28.4% (n=27) χ2(1)=0.50, p=0.48

Valproate 37.5% (n=18) 15.8% (n=15) χ2(1)=8.47, p=0.004

Atypical Antipsychotics 37.5% (n=18) 28.4% (n=27) χ2(1)=1.22, p=0.27

Typical Antipsychotics 4.2% (n=2) 7.4% (n=7) Fisher’s Exact p=0.36

Carbamazepine 2.1% (n=1) 5.3% (n=5) χ2(1)=0.80, p=0.37

Lamotrigine 12.5% (n=6) 15.8% (n=15) χ2(1)=0.28, p=0.60

Antidepressants 47.9% (n=23) 49.5% (n=47) χ2(1)=0.031, p=0.86

Benzodiazepines/ Sedatives 27.1% (n=13) 29.0% (n=27) χ (1)=0.059, p=0.81

Characteristic

CKD Cases

(n=21,741)

Non-CKD Controls

(n=86,930)

Age at index date

70-74 5,029 (23.1%) 20,157 (23.2%)

75-84 11,759 (54.1%) 47,063 (54.1%)

85+ 4,953 (22.8%) 19,710 (22.7%)

Female Gender 11,956 (55.0%) 47,808 (55.0%)

Diabetes Mellitus 10,796 (49.7%) 24,606 (28.3%)

Hypertension 20,122 (92.6%) 68,714 (79.0%)

Ischemic Heart Disease 8,217 (37.8%) 17,446 (20.1%)

Acute Kidney Injury 896 (4.1%) 447 (0.5%)

Nephrogenic Diabetes Insipidus 372 (1.7%) 851 (1.0%)

Medication Use in the 5 years prior to index date

Angiotensin-converting enzyme Inhibitors (ACEIs) 12,984 (59.7%) 35,969 (41.4%)

Angiotensin Receptor Blockers ARBs 9,069 (41.7%) 22,853 (26.3%)

NSAIDs or COX2-Inhibitors 13,406 (61.7%) 48,896 (56.2%)

Hydrocholorothiazide 10,762 (49.5%) 30,952 (35.6%)

Loop diuretics 8,318 (38.3%) 13,805 (15.9%)

Potassium-sparing diuretics 3,674 (16.9%) 6,583 (7.6%)

Typical antipsychotics 1,755 (8.1%) 4,665 (5.4%)

Atypical antipsychotics 1,210 (5.6%) 3,736 (4.3%)

Antidepressants 9,587 (44.1%) 32,982 (37.9%)

Lamotrigine 38 (0.2%) 110 (0.1%)

Carbamazepine 302 (1.4%) 1,067 (1.2%)

Benzodiazepines or Zopiclone 9,956 (45.8%) 35,612 (41.0%)

Baseline Characteristics – Lithium

Incident CKD

*Unadjusted

OR [95%CI]

Incident CKD

**Adjusted

OR[95%CI]

P-value

1) Duration of Lithium only (n=529)

< 1 year 1.10 [0.66-1.82] 1.11 [0.63-1.95] 0.72

1-2 years 0.89 [0.39-2.01] 1.25 [0.52-3.01] 0.63

2-3 years 0.74 [0.33-1.67] 0.73 [0.31-1.75] 0.48

3-4 years 1.49 [0.87-2.55] 2.28 [1.25-4.13] 0.007

>4 years 1.62 [1.26-2.10] 2.22 [1.67-2.95] <0.0001

2) Duration of Valproate only (n=498)

< 1 year 1.10 [0.77-1.57] 0.85 [0.58-1.27] 0.44

1-2 years 1.16 [0.57-2.33] 0.89 [0.42-1.87] 0.76

2-3 years <0.5 [0.09-1.55] 0.31 [0.07-1.37] 0.12

3-4 years 1.54 [0.85-2.79] 1.06 [0.55-2.04] 0.87

>4 years 1.36 [0.98- 1.87] 1.37 [0.96-1.96] 0.083

3) Previous use of both Lithium and Valproate (n=91)

N/A 1.98 [1.28- 3.06] 1.95 [1.19-3.18] 0.008

4) No Lithium, No Valproate (Reference, n=107,553)

N/A 1.00 1.00 -

Other Risk FactorsParameter

Incident CKD Unadjusted

Incident CKD *Adjusted

OR 95%CI OR 95%CIDiabetes 2.53 2.46 2.61 1.82 1.76 1.88Hypertension 3.34 3.17 3.53 1.54 1.45 1.63Ischemic Heart Disease 2.47 2.39 2.55 1.60 1.54 1.66Nephrogenic Diabetes Insipidus (NDI) 1.77 1.56 2.00 1.55 1.36 1.78Acute Kidney Injury (AKI) 8.47 7.54 9.52 6.40 5.63 7.27Loop Diuretics 3.41 3.29 3.53 2.34 2.25 2.43Hydrochorothiazide 1.79 1.74 1.85 1.37 1.32 1.42NSAIDs/Cox-2 Inhibitors 1.26 1.22 1.29 1.13 1.09 1.17ACEI/ARBs 3.47 3.34 3.60 1.85 1.77 1.93Atypical Antipsychotics 1.32 1.23 1.41 1.16 1.07 1.25Statins 1.94 1.88 2.00 1.07 1.03 1.11Lithium only 1.37 1.12 1.67 1.76 1.41 2.19Valproate only 1.21 0.98 1.49 1.03 0.81 1.29Mixed Lithium/Valproate 1.98 1.28 3.06 1.95 1.19 3.18

Neither Lithium nor Valproate use (reference) 1.00 1.00