Self-sampling for improved longitudinal monitoring of lithium treatment

Therapeutic drug monitoring is essential for safe and effective lithium treatment in bipolar disorder given its narrow therapeutic window and risk of toxicity. However, repeated venous sampling remains a barrier to use in routine care despite the well documented efficacy as a mood stabilizer. Monitoring requirements (typically every 3–6 months) are frequently not met in clinical practice, and are further challenged in low- and middle-income settings where laboratory access is limited. Even in high-income countries, monitoring is often less frequent than guideline recommendations, which may contribute to lithium toxicity and renal adverse effects. Alternative sampling strategies are therefore needed, particularly approaches that can be performed reliably outside clinical settings without compromising analytical accuracy. We therefore evaluated a patient-centric approach using volumetric dried blood spot (qDBS) sampling for longitudinal lithium monitoring.

In this prospective study, 36 adults with bipolar disorder were recruited; 31 participants contributed 117 paired serum and qDBS samples. Over four visits (mean follow-up ~60 days), venous serum lithium and capillary qDBS samples were collected 12 hours post-dose. After initial nurse-assisted sampling, participants performed supervised self-sampling. Lithium concentrations were quantified using atomic absorption spectrometry (qDBS) and routine clinical assays (serum). Agreement and correlation between methods were assessed using regression and correlation analyses. Baseline serum lithium and qDBS measurements were used to derive an individual calibration ratio, applied to estimate serum lithium at subsequent visits.

Mean serum lithium concentration was 0.65 mmol/L, with a mean qDBS-to-serum ratio of 0.83. qDBS measurements showed strong agreement with serum values (r = 0.88; R² = 0.78) and minimal bias (mean absolute difference 0.02 mmol/L). Prediction accuracy of estimated serum lithium was high (mean absolute percentage error 8.3%) and improved when two baseline measurements were used (r = 0.92).Hematocrit and clinical covariates had minimal impact on measurement performance, and no relevant effects of sampling assistance or storage time were observed.

Samples from seven individuals originally collected in 2021 were re-analyzed, enabling assessment of calibration stability over a three-year interval. In this subset, calibration appeared stable, suggesting potential long-term applicability.

A patient questionnaire indicated high acceptability of self-sampling, with usability, practicality, and perceived impact on adherence and motivation rated 8–9/10. Participants generally reported that the procedure became easier after initial instruction and was feasible for repeated use.

These findings demonstrate that individually calibrated qDBS sampling is a reliable method for lithium monitoring in bipolar disorder while being well accepted by patients. This approach may reduce dependence on venous sampling and enable decentralized or home-based therapeutic drug monitoring. Implementation could improve adherence to monitoring guidelines, support dose optimization, and enhance long-term safety of lithium treatment in routine psychiatric care, while improving accessibility and reducing healthcare burden associated with frequent laboratory-based monitoring, especially in healthcare systems where access to phlebotomy and laboratory infrastructure is limited or unevenly distributed.

References

No references