Abstract
Aim We aimed to review cases of Syndrome of Irreversible Lithium-Effectuated Neurotoxicity (SILENT) characterized by neurological sequelae following acute lithium toxicity and to explore whether cerebellar sequelae are more frequent in cases presenting with fever and/or infection.
Methods Case reports were identified from: (i) 6 reviews published up to 2005; (ii) MEDLINE, Web of Sciences, Cochrane Library and PsycINFO search.
Results We identified 123 SILENT cases published from 1965 to 2019, in which cerebellar sequelae were observed in an overwhelming proportion (79%). Nearly two out of three cases (63%) had maximal lithium plasma level <2.5 mEq/l (low/mild toxicity). Fever and/or infection were reported in nearly half of the patients (48%). The likelihood of presenting with cerebellar vs. other neurological sequelae was independently increased by elevated plasma lithium level (≥ 2.5 mEq/l) (OR=4.36, 95%CI 1.31-14.52, p = 0.02) and by a history of fever and/or infection (OR=6.48, 95%CI 2.0-21.0, p = 0.002).
Conclusions During the SARS-CoV-2 pandemic, prescribers have to be aware of the risks of cerebral sequelae associated with infection and fever in lithium users, and should warn them of the need to consult in case of fever to adjust their lithium dosage. As the occurrence of SILENT is exceptional, there is no need to modify lithium treatment preventively because of the pandemic as the benefit/risk balance of this drug remains largely positive.
Key Words: lithium/ adverse drug reaction / neurotoxicity / fever / infection

1. Introduction

Lithium has well-established benefits regarding mood stabilization and prevention of suicide, which most often outweigh the risks associated with its use [1-3]. Intoxication is the most dreaded complication of lithium exposure since it is potentially life-threatening [4-7]. Exceptionally, lithium poisoning may be complicated by the occurrence of irreversible neurological sequelae. First described by Verbov in 1965 [8], this complication was defined by Schou in 1984 as the persistence of neurological sequelae more than two months after lithium discontinuation [9]. In 1987, Adityanjee proposed to name this complication SILENT (Syndrome of Irreversible Lithium-Effectuated NeuroToxicity) [10]. Although a wide range of persisting neurological syndromes have been reported, a consistent finding across the literature is the high proportion of cerebellar sequelae, especially ataxia and dysarthria [9, 11-15]. The occurrence of neurological sequelae in lithium users is a dramatic event as limited therapeutic resources are available (physical rehabilitation, speech therapy and cognitive training) [11, 13].
The most well-documented characteristic of SILENT is that it may occur in persons with plasma lithium levels within the therapeutic range, and at any time during lithium treatment [9, 11-15]. The potential neurotoxic impact of antipsychotic-lithium polytherapy has been widely cited in the literature on the risk factors of SILENT since the publication of four cases occurring in persons treated with haloperidol and lithium [16]. The possible role of fever and infections in the occurrence of cerebellar sequelae was first mentioned by Schou [9]. This hypothesis is supported by the high proportion of persons presenting with infections and/or fever among SILENT cases [9, 11-15].
Due to its narrow therapeutic index of lithium, prescribers’ require updated accurate knowledge about the medical and lifestyle conditions increasing the risk of its adverse effects. At the onset of the SARS-CoV-2 pandemic, at a time when every psychiatrist around the world had to assess the benefit/risk ratio of psychotropic drugs in persons with COVID-19, we were struck by the fact that very few psychiatrists were aware that febrile infections might increase the risk of irreversible neurological sequelae in lithium users. Since our last review on this issue was undertaken nearly 30 years ago [15], we suspected that our knowledge about this complication might be outdated, motivating the current reappraisal of the literature. Of the five other reviews on SILENT cases [9, 11-14], the last one performed by Adityanjee included 90 cases published up to 2002 [11]. As antipsychotic prescribing practices have changed dramatically in lithium users since the introduction of second-generation antipsychotics [17], it is of interest to explore whether SILENT cases are still occurring. Furthermore, all prior reviews were exclusively narrative and did not explore which putative risk factors were independently associated with neurological outcome. Hence, the role of fever in the occurrence of cerebellar sequelae needs to be clarified.
The aim of the present study was to review published SILENT cases and to explore whether the occurrence of cerebellar neurological sequelae is more frequent in SILENT cases presenting with fever and/or infection, independently from other characteristics.

2. Methods

2.1. Identification of cases and search strategy
This review was conducted according to the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [18]. First, we considered the case reports identified in the six reviews published between 1983 and 2005 on neurological sequelae following lithium toxicity [9, 11-15]. Second, in order to identify new case reports not included in these prior reviews, we performed a MEDLINE, Web of Sciences, Cochrane Library and PsycINFO search from inception through July 2020 using the term “lithium” in combination with the following search terms (neurotox* OR neurologic* OR cerebel* OR SILENT). We examined related references of selected papers. Titles and abstracts of retrieved citations were screened, selected full-text articles were assessed independently for eligibility and data were extracted independently by two researchers. Disagreement was resolved by discussion.
We considered articles meeting the following inclusion criteria: (i) published in English, French, Spanish, Portuguese, Italian or German in peer-reviewed journals; (ii) reporting cases of neurological sequelae persisting more than two months after lithium discontinuation [9, 10]. We did not consider cases for which the observation period after lithium discontinuation was not specified or lasted less than two months, including cases with fatal outcome within the two-month period.
For each case, we extracted the following information: (i) reference: first author’s name, journal, year of publication, (ii) demographic characteristics: age, gender; (iii) characteristics of lithium treatment: dose (mg/day), maximum plasma level (mEq/l); (iv) co-prescribed drugs: antipsychotics, other drugs; (v) associated medical conditions: fever, infection, dehydration, etc.; (vi) type of neurological sequelae: cerebellar (irrespective of presence of other sequelae) vs other (any other type of sequelae without associated cerebellar sequelae). Lithium maximum plasma level (mEq/l) was categorized according to the criteria proposed by the Extracorporeal Treatments in Poisoning (EXTRIP) Workgroup [19]: i) no toxicity: <1.5 mEq/l); (ii) mild toxicity: [1.5-2.5[ mEq/l; (iii) moderate toxicity: [2.5–3.5[ mEq/l; (iv) severe toxicity: > 3.5 mEq/l.
2.2. Statistical analyses
Multivariate logistic regressions giving Odds Ratio and 95% confidence intervals (OR, 95% CI) were used to identify the characteristics independently associated with the type of neurological sequelae (cerebellar vs. other). Age was categorized as < 50 vs. ≥ 50 ys. Lithium maximum plasma level (mEq/l) was dichotomized as “no/mild” (<2.5 mEq/l) vs. “moderate/severe” (≥ 2.5 mEq/l) toxicity. Antipsychotic exposure was categorized as present vs. absent irrespective of the number and type of antipsychotics. Exposure to fever and/or infection was defined by the presence of at least one of the two conditions, irrespective of the cause of fever and the site of infection. Strictly defined fever included only cases with reported information on this symptom, irrespective of the cause. All the variables (age, gender, lithium level, antipsychotic use, fever) were simultaneously entered in the regression model. The models were not adjusted for lithium dose as this information was missing in a large proportion of cases.

3. Results

3.1. Results of literature search
Figure 1 presents a flow chart of the eligibility process for this review. One hundred cases published from 1965 to 2002 were identified in the 6 prior reviews [9, 11-15]. The numbers of additional cases fulfilling the inclusion criteria identified in each review are given in the flow chart. The systematic search identified 23 additional cases published from 1986 to 2019 [20-39].
Ultimately, we included 123 cases of neurological sequelae associated with lithium exposure reported in 92 articles published from 1965 to 2019 (Table 1). The number of reported cases peaked in the ‘80s. More than half of the 123 cases were published at the end of this decade, while very few cases have been published since the beginning of the 21st century (1960s: n=1; 1970s: n=21, 1980s: n= 48; 1990s: n=30; 2000s: n=15; 2010s: n=8). Most articles (n=75, 81%) reported a single case, and the maximal number of cases in an article was 7 [13].
3.2. Demographic characteristics
There were 53% (n=65) females and 47% (n=58) males. The mean age of cases was 46.8 (standard deviation 13.3, range 20-70), with 57 cases (46.3%) aged ≤ 50 years.
3.3. Lithium treatment
The median daily prescribed dose of lithium was 1200 mg (interquartile range 900-1500). Twelve (9.8%) overdoses were reported (intentional n=8, accidental n=3; not specified n=1).
Using the EXTRIP classification, the distribution of maximum lithium level was: (i) no toxicity <1.5 mEq/l n= 42 (38.5%); (ii) mild toxicity [1.5-2.5[ mEq/l n=27 (24.8%); (iii) moderate toxicity [2.5–3.5[ mEq/l n=20 (18.4%); (iv) severe toxicity > 3.5 mEq/l n=20 (18.4%). A plasma level > 5 mEq/L indicating the need for extracorporeal treatments (ECTRs) [19] was reported in 7 (6.4%) of patients. Information on lithium level was missing in 24 cases.