Response to ‘Isotretinoin and adverse neuropsychiatric outcomes: retrospective cohort study using routine data’

Abstract

Dear Editor, We read with great interest the study by Paljarvi et al., which indicated that isotretinoin was not associated with longstanding concerns about the risk of neuropsychiatric outcomes, and even more so, that it did reduce the neuropsychiatric risk in patients with moderate‐ to‐severe acne.¹ We thank the authors for conducting this excellent study by collecting population‐based electronic medical records from TriNetX, and for pointing out a practical suggestion. However, we would like to share ideas on this article.

Firstly, the authors conducted comprehensive propensity score matching for many covariates at baseline, including mental and behavioural disorders; psychotic disorders; mood disorders; anxiety, dissociative, stress‐related, somatoform and other nonpsychotic mental disorders; adult personality and behavioural disorders; typical behavioural and mood disorders of childhood and adolescence; sleep disorders; and self‐harm; etc. However, many of these covariates were also the primary outcomes of interest measured by the authors. Would it be possible for the authors to perform a sensitivity analysis to exclude patients with these covariates from the cohort entry date to avoid collinearity² during statistical analysis and make their results more reliable?

Secondly, this article provided a well‐defined model for retrospective cohort studies. The authors set a 14‐day washout period after the index prescription date to mitigate possible reverse causality, and calculated odds ratios. We suggest the authors use a proportional hazards model (a built‐in function of TriNetX) that takes into account the time to events to depict the risk in a time trend. It is of merit for readers to know the incidence of events during a specific time interval after exposure to isotretinoin.

Thirdly, although Paljarvi et al. compared the incidence of adverse neuropsychiatric outcomes under the use of isotretinoin with other populations, two more issues about the drug safety need further evaluation. One is the association between treatment duration of isotretinoin and the risk of a neuropsychiatric adverse event. Another is the dose relationship between the use of isotretinoin and these adverse neuropsychiatric outcomes.

In conclusion, Paljarvi et al. have provided robust evidence about the association between isotretinoin and adverse neuropsychiatric outcomes. However, safety issues remain the major concerns in the use of isotretinoin.³ Further analysis is warranted.

Author contributions

Renin Chang: Conceptualization (equal); writing – original draft (lead); writing – review and editing (supporting). Chih‐Cheng Lai: Conceptualization (equal); writing – review and editing (equal). Yao‐Min Hung: Methodology (equal); writing – review and editing (equal). Shiow‐Ing Wang: Methodology (equal); validation (equal). James Cheng‐Chung Wei: Methodology (equal); supervision (equal); writing – review and editing (equal).

References

Author notes