Skip to main content
. 2021 May 25;144(3):259–276. doi: 10.1111/acps.13312

TABLE 2.

Combination of lithium and valproic acid

References № of patients Certainty assessment Effect
Combination Comparison treatment Study design Outcome/Event Follow‐up time Risk of bias Inconsistency Indirectness Imprecision Other considerations Relative (95% CI) Absolute (95% CI) Test (p‐value) Certainty
Valproic acid + Lithium Valproic acid
The BALANCE investigators and collaborators, 201013

110

110

69.0%

Randomized trials Time to intervention for any mood episode 24 months Seriousa Not serious Not serious Not serious None

HR 0.57

(0.40–0.80)

−20%

(−32 to −8)

Log‐rank (0.0014)

⨁⨁⨁◯

MODERATE

The BALANCE investigators and collaborators, 201013

110

110

45.0%

Randomized trials Time to intervention for depressive episode 24 months Seriousa Not serious Not serious Seriousb None

HR 0.70

(0.46–1.07)

−11%

(−21 to +2)

Log‐rank (0.11)

⨁⨁◯◯

LOW

The BALANCE investigators and collaborators, 201013

110

110

45.0%

Randomized trials Time to intervention for manic episode 24 months Seriousa Not serious Not serious Not serious None

HR 0.51

(0.32–0.80)

−19%

(−28 to −7)

Log‐rank (0.003)

⨁⨁⨁◯

MODERATE

The BALANCE investigators and collaborators, 201013

110

110

68.0%

Randomized trials Time to new drug treatment 24 months Seriousa Not serious Not serious Not serious None

HR 0.57

(0.40–0.80)

−20%

(−31 to −8)

Log‐rank (0.0014)

⨁⨁⨁◯

MODERATE

The BALANCE investigators and collaborators, 2010 13

110

110

23.0%

Randomized trials Time to hospital admission 24 months Seriousa Not serious Not serious Not serious None

HR 0.51

(0.27–0.96)

−11%

(−16 to −1)

Log‐rank (0.038)

⨁⨁⨁◯

MODERATE

The BALANCE investigators and collaborators, 201013

110

110

44.0%

Randomized trials Time to medication discontinuation 24 months Seriousa Not serious Not serious Very seriousc None

HR 1.01

(0.67–1.53)

0%

(−12 to +15)

Log‐rank (0.95)

⨁◯◯◯

VERY LOW

Ahearn EP et al., 201311 145 821 Observational studies Suicide attempts 10,000 months on drug regimen 6 years, mean 19.8 months Very serious c , d Not serious Seriouse Not serious None Valproate + Lithium: 6.3; Valproate: 7.0.

⨁◯◯◯

VERY LOW

Valproic acid + Lithium Lithium
The BALANCE investigators and collaborators, 201013

110

110

59.0%

Randomized trials Time to intervention for any mood episode 24 months Seriousa Not serious Not serious Seriousb None

HR 0.80

(0.57–1.15)

−6%

(−18 to +6)

Log‐rank (0.23)

⨁⨁◯◯

LOW

Kemp DE et al., 200914

16

15

55.0%

Randomized trials Time to intervention for any mood episode 6 months Seriousd Not serious Seriousf Very serious b , c , g None

HR 0.72

(0.32–1.65)

−11%

(−32 to +18)

Log‐rank (0.44)

⨁◯◯◯

VERY LOW

Wingard L et al., 201912

217

1133

87.0%

Non‐randomized studies Time to treatment failure 365 days Very serioush Not serious Not serious Not serious None

HR 0.72

(0.62–0.85)

−10%

(−15 to −5)

⨁⨁◯◯

LOW

The BALANCE investigators and collaborators, 201013

110

110

32.0%

Randomized trials Time to intervention for depressive episode 24 months Seriousa Not serious Not serious Very serious b , c None

HR 1.06

(0.67–1.67)

+2%

(−9 to +15)

Log‐rank (0.81)

⨁◯◯◯

VERY LOW

The BALANCE investigators and collaborators, 201013

110

110

36.0%

Randomized trials Time to intervention for manic episode 24 months Seriousa Not serious Not serious Seriousb None

HR 0.66

(0.41–1.07)

−11%

(−15 to +2)

Log‐rank (0.09)

⨁⨁◯◯

LOW

The BALANCE investigators and collaborators, 201013

110

110

20.0%

Randomized trials Time to hospital admission 24 months Seriousa Not serious Not serious Very serious b , c None

HR 0.72

(0.38–1.38)

−5%

(−12 to +7)

Log‐rank (0.32)

⨁◯◯◯

VERY LOW

Kemp DE et al., 200914

217

1133

20.4%

Non‐randomized studies Time to hospital admission 365 days Very serioush Not serious Not serious Not serious None

HR 1.05

(0.78–1.40)

−1%

(−41 to +69)

⨁⨁◯◯

LOW

The BALANCE investigators and collaborators, 201013

110

110

45.0%

Randomized trials Time to medication discontinuation 24 months Seriousa Not serious Not serious Very serious b , c None

HR 0.89

(0.60–1.34)

−4%

(−15 to +10)

Log‐rank (0.58)

⨁◯◯◯

VERY LOW

Kemp DE et al., 200914

217

1133

20.2%

Non‐randomized studies Time to medication discontinuation 365 days Very seriouse Not serious Not serious Not serious None

HR 0.63

(0.44–0.89)

−7%

(−11 to −2)

⨁⨁◯◯

LOW

The BALANCE investigators and collaborators, 201013

110

110

58.0%

Randomized trials Time to new drug treatment 24 months Seriousa Not serious Not serious Seriousb None

HR 0.80

(0.56–1.14)

−8%

(−20 to +5)

Log‐rank (0.21)

⨁⨁◯◯

LOW

Kemp DE et al., 200914

217

1133

46.5%

Non‐randomized studies Time to new drug treatment 365 days Very serioush Not serious Not serious Not serious None

HR 0.62

(0.49–0.79)

−14%

(−20 to −8)

⨁⨁◯◯

LOW

Solomon DA et al., 199716 0/5 (0.0%) 5/7 (71.4%) Randomized trials Relapse frequency 12 months Not serious Not serious Seriousi Very serious c , g None

RR 0.12

(0.01–1.79)

−63%

(−71 to +56)

Chi‐squared (0.014)

⨁◯◯◯

VERY LOW

Ahearn EP et al., 201311 145 436 Observational studies Suicide attempts 10,000 months on drug regimen 6 years, mean 19.8 months Very seriousc,d Not serious Seriousj Not serious None Valproate + Lithium: 6.3; Lithium: 7.7.

⨁◯◯◯

VERY LOW

Valproic acid + Lithium CBZ + Lithium
Missio G et al., 201915 33 27 Randomized trials Time to relapse 22 weeks Very serious a , d , k Not serious Not serious Seriousg None Time‐to‐relapse curves were not significantly different between groups Log‐rank (0.45)

⨁◯◯◯

VERY LOW

CBZ, Carbamazepine; CI, Confidence interval; HR, Hazard Ratio; MD, Mean difference; RR, Risk ratio. Explanations:

a

Open‐label trial.

b

Low power for the outcome.

c

Large CI including significantly reduced and increased effect.

c

No measures of disease severity or comorbidities available.

d

Incomplete outcome reporting.

e

Days of prescription used as proxy of days of treatment.

f

Population of rapid cycling patients with recent alcohol / illicit drug abuse or dependence.

g

Small sample size.

h

Case severity assessed by length of index hospitalization and presence of psychotic symptoms. Lack of concealment for allocation.

i

4 patients never recovered from the index episode, but they were equally distributed between the treatment arms.

j

Days of prescription used as proxy of days of treatment.

k

No definition of relapse.