The effect of lithium variation coefficient on the risk of attack in patients with bipolar disorder: A pilot study

Abstract

Background

This study examines the association between the coefficient of variation (%CV) of lithium levels and episode risk and frequency in bipolar patients maintaining serum lithium levels within the therapeutic range.

Methods

We retrospectively reviewed patients with bipolar disorder under care from 2018 to 2022. Inclusion criteria were at least 2 years of follow-up, a minimum of three annual lithium level measurements within the therapeutic range. Patients were categorized based on seizure status. We calculated mean lithium levels, standard deviation (SD), and %CV.

Results

The study included 75 patients (patients with-without episodes, 39–36). Demographic data revealed no significant differences. While mean lithium levels showed no significant disparity between groups, SD and %CV were notably higher in patients with episodes (P < .05). ROC analysis demonstrated AUC values of 0.722 (95% CI: 0.607–0.836 P = .001) for %CV and 0.709 (95% CI: 0.593–0.826; P = .002) for SD. The optimal %CV cutoff was 17.39, with 67% sensitivity and 69% specificity. A weak correlation was found between %CV and the number of episodes (P = .001, r = 0.376). The post-hoc power analysis for this study was 0.78.

Conclusions

Despite acceptable lithium levels, patients with recent episodes exhibited significant lithium level fluctuations. Integrating %CV with real-time lithium measurements during bipolar disorder follow-up may enhance clinical monitoring and seizure prediction.

Keywords

Bipolar disorder lithium coefficient of variation

Introduction

Bipolar disorder is a psychiatric disorder in which episodes of manic and depressive mood recur with different clinical courses from person to person. Its prevalence in the general population ranges from 1% to 3%, but this can increase to 4% to 15% in individuals with a family history of bipolar disorder.¹ In patients diagnosed with bipolar disorder, regular treatment and follow-up are required to reduce the frequency of episodes, manage the psychological and social effects of the disease, and increase the person’s functionality. The primary approach to treatment involves pharmacotherapy, often complemented by psychotherapy.² Continuity and follow-up of treatment are important because 70% of patients have a risk of relapse within 5 years despite treatment. Available treatment options include single or combination therapy with lithium, valproic acid, lamotrigine, carbamazepine, antipsychotics, and benzodiazepines.³ According to current treatment guidelines, the therapeutic range of lithium, which is still the first choice in treating bipolar disorder, should be 0.6–1.5 mmol/L in patient serum.⁴ However, this drug has a narrow therapeutic window, with potential side effects like sedation, tremors, weight gain, diarrhoea, nausea, polydipsia, polyuria, cognitive disturbances, hypothyroidism, and diabetes insipidus. Fatal doses can result from uncontrolled use. Hence, serum lithium levels should be monitored every 1–2 weeks following the initiation of the medication and every 3–6 months for regular follow-up.⁵ Despite these side effects, lithium effectively reduces the risk of recurrence of mood episodes.⁶ Individual characteristics such as age and physical attributes can also influence lithium levels.⁷

The coefficient of variation (%CV) is a statistical value that makes it possible to calculate the variability of simple, classified data series with different means, regardless of the unit of measurement used for these numbers and the mean of the numbers. It is obtained by dividing the standard deviation of the data set by its arithmetic mean, then multiplying the result by 100. The %CV is expressed as a percentage, and it does not have a specific unit of measurement.⁸

Studies in the literature examine the relationship between patients’ instantaneous serum lithium levels and the risk of an attack.^9,10 Regular lithium use protects patients with bipolar disorder from attacks.¹¹ In clinical observations, patients have been reported to experience manic and/or depressive episodes even if serum lithium levels are in the therapeutic range.¹² We observed that some patients with bipolar disorder had attacks despite serum lithium levels in the therapeutic range. As far as we know, there are studies in the literature related to serum lithium levels, but no study conducts variation analysis. The aim of this study is to assess the fluctuation in lithium levels with the %CV among patients diagnosed with bipolar disorder who experience episodes despite having serum lithium levels within the therapeutic range during follow-up. Additionally, we aim to investigate the correlation between these fluctuations and the risk and frequency of episodes.

Materials and methods

The study has been approved by Adiyaman University Institutional Clinical Research Ethics Committee (Approval No: 2022/9-6) and conducted following the Helsinki Declaration. The diagnosis of bipolar disorder was made according to the Fifth Edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5).¹³ The clinical notes and laboratory results of bipolar disorder patients who presented to the psychiatry clinic between 2018 and 2022 were retrospectively reviewed in the hospital information system. Patients who were followed for at least 2 years, had at least three blood lithium levels per year, and had lithium levels within the therapeutic range (0.6–1.5 mmol/L) during their follow-up were included in the study. Patients with additional organic diseases, such as psychiatric or renal conditions, which could influence blood lithium levels, as well as those with fewer than three blood lithium results per year, and those without regular follow-up, were excluded from the evaluation. The patients included in the study were evaluated into two groups, based on their examination notes: those who experienced an episode in the last 2 years of follow-up and those who remained without episode for 2 years. Manic episode considered as a distinct period of abnormally and persistently elevated, expansive, or irritable mood, lasting at least 1 week and be present most of the day, nearly every day. During the period of mood disturbance, three (or more) of the following symptoms have persisted (four if the mood is only irritable) and have been present to a significant degree: Inflated self-esteem or grandiosity, decreased need for sleep, increased talkativeness, racing thoughts, distractibility, increase in goal-directed activity or psychomotor agitation, excessive involvement in pleasurable activities that have a high potential for painful consequences. And for the diagnosis of depressive episode, a patient must experience five or more of the following symptoms, which must persist most of the day, daily, for at least 2 weeks in a row: depressed mood, loss of interest or pleasure in all/almost all activities, significant weight loss or decrease or increase in appetite, sleep disturbance, fatigue or loss of energy, feelings of worthlessness or guilty, recurrent thoughts of death or suicidal ideation-attempts, diminished ability to think or concentrate or indecisiveness, severe psychomotor changes (agitation or retardation).¹³ At our center’s Mental Health and Diseases Clinic, blood lithium levels are checked at 3–6 month intervals in bipolar disorder patients in remission. Patients have 5 cc of venous blood drawn 12 h after the nighttime lithium dose. Venous blood samples are obtained using the antecubital vein. The samples are analysed using the spectrophotometric method on the Abbott Architect c16000 analyser (Abbott Diagnostics Inc., Lake Forest, IL, USA), and the results for patients are uploaded to the patient record system within a maximum of 3 days. According to this device, the reference range for lithium is 0.6–1.5 mmol/L.

The study data were analysed using Statistical Package for Social Sciences (SPSS) for 23.0 Windows software programs, and a significance level of 0.05 was considered. The mean, standard deviation (SD), and CV% of the 2-year lithium results for each patient included in the study were calculated. The formula ‘(SD/mean) × 100’ was used for %CV. The Kolmogorov–Smirnov test assessed the deviation from a normal distribution. Quantitative variables were summarized as median (interquartile range [IQR]) for non-parametric data. Non-parametric continuous data were analysed by Mann–Whitney’s U tests. Spearman’s Rank test did the analysis of correlation. Receiver-operating characteristics (ROC) analysis was performed to estimate the sensitivities and specificities of %CV to evaluate attack risk. A statistical post-hoc power analysis was done by G-power 3.1.9.2 software program.¹⁴

Results

Between 2018 and 2022, a total of 1285 patients presented to the hospital with a diagnosis of bipolar disorder. After applying exclusion criteria, a total of 75 patients, including 39 in the patients with episodes and 36 in the patients without episode, were included in the study. Of the all patients included in the study, 48% (n = 36) were female, with a median age of 43 years (33–46). No statistically significant difference was found between the patients with episodes and without episode regarding age and gender (sequentially; P = .857 ve P = .554). The demographic data for both patient groups are given separately in Table 1.

Table 1.

The demographic data of the patients and the mean, SD, and %CV values for the 2-year lithium results are as follows.

	Patients with episodes (n = 39)	Patients without episodes (n = 36)	P value
Age^a, year	42 (33–46)	43 (34.25–46)	0.857 (z = −0.181)
Gender
• Man, n (%)	19 (48.7%)	20 (55.6%)	0.554 (χ2 = 0.351)
• Woman, n (%)	20 (51.3%)	16 (44.4%)
Duration of illness^a, year	10 (6–12.5)	10 (8–16.5)	0.311 (z = −1.013)
Type of treatment
• Only lithium, n (%)	31 (79.5%)	24 (66.7%)	0.210 (χ2 = 1.573)
• VA + lithium, n (%)	8 (20.5%)	12 (33.3%)
Recent episode type
• Manic, n (%)	24 (61.5%)
• Depressive, n (%)	15 (38.5%)
Number of episodes^a, in the last 2 years	2 (1–3)	0
For the six lithium results for each individual patient
• Mean^a, mmol/L	0.78 (0.70–0.88)	0.76 (0.72–0.80)	0.408 (z = −0.827)
• SD^a, mmol/L	0.15 (0.12–0.22)	0.12 (0.06–0.15)	0.002 (z = −3.118)
• CV%^a	20.13 (15.69–26.5)	15.17 (9.46–19.43)	0.001 (z = −3.298)

^aThe data is presented as median (Interquartile range). %CV: Coefficient of variation, SD: Standard deviation, VA: Valproic acid.

The mean and SD values of the last six lithium results within the previous 2 years for patients in both groups, as well as the %CV values calculated using these two parameters, are provided in Table 1. Among these both groups with therapeutic range lithium values in follow-up, no significant difference was detected between the mean lithium levels. In contrast, the SD and %CV values were significantly higher in the patients with episodes (P < .05). In the subgroup analysis, where we grouped the patients with episodes group based on the last attack type, there was no statistically significant difference between those with the previous attack being manic and those with the latest attack being depressive in terms of mean, SD, and %CV values.

In the ROC analysis conducted to determine the superiority of SD and %CV in predicting the attack, the area under the ROC curve was found to be 0.709 (95% CI 0.593–0.826; P = .002) for SD and 0.722 (95% CI 0.607–0.836; P = .001) for %CV. According to the calculation based on the Youden index, the optimal value of %CV in predicting the attack was 17.39 or higher, with a sensitivity of 67% and specificity of 69%.

A statistically significant but weak correlation was found in the correlation analysis between the number of attacks patients experienced in the last 2 years and the %CV values of their lithium levels in the previous 2 years (P = .001, r = 0.376).

In the post-hoc power analysis using the G-power program, the effect size was 0.70 and found to be medium-large. With α = 0.05 and 0.70 effect size, the post-hoc analyses revealed that the statistical power for this study was 0.78.

Figure 1 was generated to demonstrate the application of %CV. In Figure 1, line graphs depict six lithium results over the past 2 years for three bipolar patients who experienced an episode despite having therapeutic lithium levels and three control bipolar patients who maintained therapeutic lithium levels and remained episode-free for 2 years. While %CV values were found to be high in the patients with episodes group (%CV values respectively: Patient 1, Patient 2, Patient 3; 28.5, 37.5, 29.3), it was observed that %CV values were lower in the patients without episode group due to lithium levels being within normal limits and within a narrower range in the last 2 years (%CV values respectively: Control 1, Control 2, Control 3; 4.82, 7.07, 7.10).

Figure 1.

Two-year lithium levels of patients (three from the control group and three from the patient group).

Discussion

We observed that %48 of bipolar disorder patients using lithium, both those who experienced manic/depressive episodes and those who did not, were female. Although the course and clinical characteristics of bipolar disorder may vary between women and men, there is no evidence to suggest that gender influences the response to mood stabilizers in treatment.¹⁵ We also did not detect any differences in terms of gender between bipolar disorder patients who experienced episodes and those who did not. The mean age of the patients was 43 (33–46), and we did not find any difference in mean age between the two groups. However, bipolar disorder is typically seen in younger people.¹⁶ While age can significantly affect mood-stabilizing medications’ protective effects, we believe patients tend to be more motivated for regular blood tests, follow-up, and treatment compliance as age advances.^17,18 Therefore, we observed that patients with lithium levels checked at least three times a year in the last 2 years were older.

In this study, we did not observe any difference in mean lithium levels between the groups with and without episode among individuals diagnosed with bipolar disorder and using lithium. However, we observed that individuals who experienced episodes had higher SD and %CV values regarding lithium levels. A meta-analysis study conducted with 1406 individuals reported that serum lithium levels between 0.4 and 0.8 mmol/L significantly protected bipolar patients from major mood episodes.¹⁹ However, in a prospective study where bipolar disorder patients receiving lithium monotherapy were observed for 2.5 years, it was found that patients’ mood symptoms worsened while under treatment. During the 2.5-year follow-up period, it was noted that intervention was necessary for manic or mixed mood symptoms in 27 patients, and 22 patients required treatment for depressive symptoms. It was found that the mean blood lithium levels before the recurrence of manic or mixed symptoms were lower than before the recurrence of depressive symptoms (0.53+/−0.13 vs 0.66+/−0.21 mmol/l, P = .01). In conclusion, Kleindienst et al. suggested that manic or mixed recurrences can occur at lower lithium levels, whereas depressive episodes tend to occur at higher levels.¹⁰ Severus et al. also observed that patients within the accepted therapeutic range while using lithium experienced attacks.²⁰ They noted that lithium levels in the lower range of the therapeutic spectrum may be sufficient to prevent depressive episodes, but higher lithium levels than this range may be needed for protection against manic/mixed episodes. Our study found that lithium levels within the therapeutic range and experiencing an attack did not significantly differ based on whether the last attack type was manic or depressive. It is believed that this situation may be related to the small number of cases in the study.

In the ROC analysis conducted to determine the superiority of SD and %CV in predicting attacks, we observed that the variation coefficient had a sensitivity of 68% and specificity of 67% in predicting attack risk. We observed that it may not be sufficient as a predictor. However, we believe that it can still alert clinicians to the occurrence of episodes, as observed in patient follow-up, as shown in Figure 1. %CV is superior to SD, especially in evaluating variability of simple in two groups with different mean values. In our study, although the mean values of the two groups were very close, CV was revealed to be a superior predictor in showing variability. If the means of the two groups were markedly different, CV might be expected to be an even better predictor.

Among the study’s limitations, we can mention the retrospective nature of the evaluation, which prevented the optimization of other factors that could have contributed to the patients’ episodes. Another limitation is the potential impact of drug intake and variations in blood sampling.²¹ While our patients are informed about this concern, the timing of blood collection and drug intake may vary, particularly among outpatients. Variations in %CV values could be influenced by the pharmacokinetics activity of the drug. Additionally, some patients were using valproic acid alongside lithium, and the relatively small sample size is our other limitations.

In conclusion, the coefficient of variation can be used as a parameter indicating that even in bipolar disorder patients with lithium levels within the therapeutic range, there is a high fluctuation in lithium levels over time. The high %CV values indicating lithium fluctuation may be one of the reasons behind the episodes observed in patients with lithium levels within the therapeutic range. Although weak, the observed correlation between the frequency of occasional attacks and the %CV values of lithium results supports this. During outpatient follow-ups, evaluating the patient’s laboratory results, including instant lithium levels, %CV value, can benefit the patient’s clinical assessment. We believe supporting this hypothesis with a more extensive patient series would be appropriate.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

Not applicable.

Guarantor

SBO.

Contributorship

Conceptualization [SBO, ŞS]; Data curation [SBO, ŞS]; Formal analysis [SBO, ŞS]; Investigation [SBO, ŞS]; Methodology [SBO, ŞS]; Project administration [SBO, ŞS]; Supervision [BHA, İGÇ]; Roles/Writing - original draft [SBO, ŞS, SY]; and Writing - review & editing [SBO, ŞS, SY, BHA, İGÇ].

ORCID iD

Saniye Başak Oktay

References

Miller

Hamilton

Sacchet

, et al. Meta-analysis of functional neuroimaging of major depressive disorder in youth. JAMA Psychiatr 2015; 72(10): 1045–1053.

Karaytuğ

Tamam

Demi̇rkol

, et al. Bipolar bozukluk tanılı hastalarda tedaviye uyum ve i̇lişkili faktörler. aktd 2022; 31(1): 21–27.

Cullen

Kappelmann

Umer

, et al. Efficacy and acceptability of pharmacotherapy for comorbid anxiety symptoms in bipolar disorder: a systematic review and meta-analysis. Bipolar Disord 2021; 23(8): 754–766.

Yatham

Kennedy

Parikh

, et al. Canadian network for mood and anxiety treatments (CANMAT) and international society for bipolar disorders (ISBD) 2018 guidelines for the management of patients with bipolar disorder. Bipolar Disord 2018; 20(2): 97–170.

Price

Marzani-Nissen

. Bipolar disorders: a review. Am Fam Physician 2012; 85(5): 483–493.

Geddes

Burgess

Hawton

, et al. Long-term lithium therapy for bipolar disorder: systematic review and meta-analysis of randomized controlled trials. Am J Psychiatr 2004; 161(2): 217–222.

Firth

Siddiqi

Koyanagi

, et al. The Lancet Psychiatry Commission: a blueprint for protecting physical health in people with mental illness. Lancet Psychiatr 2019; 6(8): 675–712.

Abdi

. Coefficient of variation. Encyclopedia of Research design. In: Salkind

(ed) Thousand Oaks, CA: Sage Publications, Inc., 2010. DOI: 10.4135/9781412961288.

Hsu

Tsai

Tseng

, et al. Differences in the prophylactic effect of serum lithium levels on depression and mania in bipolar disorder: a dose-response meta-analysis. Eur Neuropsychopharmacol 2022; 58: 20–29.

10.

Kleindienst

Severus

Greil

. Are serum lithium levels related to the polarity of recurrence in bipolar disorders? Evidence from a multicenter trial. Int Clin Psychopharmacol 2007; 22(3): 125–131.

11.

Volkmann

Bschor

Köhler

. Lithium treatment over the lifespan in bipolar disorders. Front Psychiatr 2020; 11: 377.

12.

Tondo

Baldessarini

Hennen

, et al. Lithium maintenance treatment of depression and mania in bipolar I and bipolar II disorders. Am J Psychiatr 1998; 155(5): 638–645.

13.

American Psychiatric Association . Diagnostic and statistical manual of mental disorders. 5th ed. Virginia: American Psychiatric Association, 2013. DOI: 10.1176/appi.books.9780890425596.

14.

Faul

Erdfelder

Lang

, et al. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 2007; 39(2): 175–191.

15.

Arnold

. Gender differences in bipolar disorder. Psychiatr Clin 2003; 26(3): 595–620.

16.

Depp

Jeste

. Bipolar disorder in older adults: a critical review. Bipolar Disord 2004; 6(5): 343–367.

17.

Özkorumak

Aykut

Tiryaki

. Comparison of moderate-to-advanced aged with young manic patients. J Mood Disord 2014; 4(2): 47–52.

18.

Colomer

Anmella

Vieta

, et al. Physical health in affective disorders: a narrative review of the literature. Br J Psychiatr 2021; 43(6): 621–630.

19.

Hsu

Carvalho

Tsai

, et al. Lithium concentration and recurrence risk during maintenance treatment of bipolar disorder: multicenter cohort and meta-analysis. Acta Psychiatr Scand 2021; 144(4): 368–378.

20.

Severus

Kleindienst

Evoniuk

, et al. Is the polarity of relapse/recurrence in bipolar-I disorder patients related to serum lithium levels? Results from an empirical study. J Affect Disord 2009; 115(3): 466–470.

21.

Reddy

. Serum lithium levels: ideal time for sample collection! Are we doing it right? Indian J Psychol Med 2014; 36(3): 346–347.