HyperCKemia and rhabdomyolysis in the neuroleptic malignant and serotonin syndromes: A literature review

Neuroleptic malignant syndrome and serotonin syndrome are two syndromes whose molecular bases remain poorly understood. The phenotypes of both syndromes overlap with other syndromes that have a clear genetic background, in particular RYR1-related malignant hyperthermia. Through a literature review, performed according to the PRISMA guidelines, we aimed to report the clinical features of both syndromes, and the results of genetic testing performed. 10 case series and 99 case reports were included, comprising 134 patients. A male predominance of 58% was found. The median age was 35 (range 4-84) years. Eight patients experienced recurrent episodes of rhabdomyolysis. Genetic analysis was performed in eleven patients (8%), revealing four RYR1 variants, three likely benign (p.Asp849Asn, p.Arg4645Gln, p.Arg4645Gln) and one variant of uncertain significance (p.Ala612Thr). This review underlines that a subset of patients with neuroleptic malignant syndrome and serotonin syndrome develop recurrent episodes of rhabdomyolysis. This recurrent pattern suggests a possible underlying (genetic) susceptibility. However, the genetic background of neuroleptic malignant syndrome and serotonin syndrome has only been investigated to a very limited degree so far. The increasing availability of next generation sequencing offers an opportunity to identify potentially associated genetic backgrounds, especially in patients with recurrent episodes or a positive family history.


Introduction
Rhabdomyolysis is a severe and potentially life-threatening medical emergency involving the dissolution of skeletal muscle cells, due to a wide range of etiologies that ultimately all converge in a common pathway leading to irreversible muscle breakdown [1] . Whilst in some instances rhabdomyolysis may be exclusively due to external triggers (e.g., trauma, local ischaemia, exertional exercise), recent observations suggest that such events more commonly reflect a synergistic effect of predisposing genetic and environmental factors [2] . Specific phenotypes elicited by pharmacological agents (malignant hyperthermia (MH), neuroleptic malignant syndrome (NMS) and serotonin syndrome (SS)) feature https://doi.org/10.1016/j.nmd.2020. 10.010 0960-8966/© 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ) muscle breakdown as an essential part of a more widely defined syndrome, and are all considered part of the rhabdomyolysis spectrum.
MH is a well described pharmacogenetic disorder which clinically manifests as a hypermetabolic crisis when an MH-susceptible (MHS) individual is exposed to volatile anesthetics ( e.g., halothane, isoflurane, sevoflurane, desflurane) or the depolarizing muscle relaxant succinylcholine. An MH reaction typically features a critically increased body temperature (often in excess of 42 °C), muscle rigidity or spasms, rhabdomyolysis, tachycardia, and other life-threatening symptoms. MH has been associated with specific mutations in the RYR1, CACNA1S or STAC3 genes, all of which ultimately result in excessive cytoplasmic calcium accumulation in response to both pharmacological and non-pharmacological triggers. To date, 48 RYR1 and two CACNA1S variants have been unequivocally classified as MH diagnostic variants ( https:// www.emhg.org/ genetics ).
In contrast to MH, the genetic background and pathophysiology of NMS or SS are relatively poorly understood. SS is a potentially fatal condition, most likely caused by high levels of serotonin (5-hydroxytryptamine, 5-HT) in the synaptic cleft, leading to an overactivation of the central and peripheral serotonin receptors. The reaction is thought to be concentration dependent, and is often the result of concomitant exposure to two or more serotonergic agents. Monoamine oxidase inhibitors, selective serotonin reuptake inhibitors, and serotonin-norepinephrine reuptake inhibitors are the medications most commonly associated with SS [3] . Simultaneous prescription of medication inhibiting the cytochromes P450 (CYPs) is associated with increased serum concentration of serotonergic drugs, leading to potentially toxic levels. Similarly, genetic polymorphisms in these CYP enzymes, in particular CYP2D6, CYP3A4, and CYP2C19, increase the serum concentration of serotonergic medication and thus the risk of reaching a potentially toxic serum concentration [4] . In addition, it appears that polymorphisms in the 5-HT-2A receptor may predispose to the development of SS [5] .
NMS is most likely caused by impaired function of central dopamine pathways [6] . This hypodopaminergic theory is based on the observation that NMS typically develops after exposure to dopamine antagonists, or acute withdrawal of dopamine agonists, referred to as the neuroleptic malignant-like syndrome (NMLS) [7] . NMS is a rare but potentially life-threatening reaction to almost any of a group of antipsychotic drugs or major tranquilizers (neuroleptics). A genetic basis for the disorder had been suspected but has not been proven yet; a case series of identical twins, and a mother and two daughters all experiencing NMS, support this suspicion of an underlying genetic risk factor [8] .
Some authors have suggested that NMS may be genetically related to MH [9][10][11] , taking into account that the clinical triad of autonomic dysfunction, altered mental status and rhabdomyolysis is seen in both MH and NMS, but also SS. The hypothesis of RYR1 involvement in NMS and SS is also supported by the observation that other similar (non-anesthesia) related clinical syndromes with rhabdomyolysis such as exertional heat stroke (EHS) and exertional rhabdomyolysis (ERM) have also been associated with (MH-related) RYR1 variants [12] . The overlap in clinical presentation and pathophysiological pathways suggests that there might be similarities between the genetic background underlying MH/ERM/EHS and NMS/SS.
In the present study we aim to review the clinical features of patients with hyperCKemia or rhabdomyolysis events in the context of NMS and SS. In addition, we aim to describe genetic testing (where performed) in particular in patients with clinical features suggestive for a monogenic background (i.e. , recurrent episodes or a positive family history).

Methods
A systematic review was performed according to the guidelines of "Preferred Reporting Items for Systematic Reviews and Meta-Analyses" (PRISMA) [13] . Methods of the analysis and inclusion criteria were specified in advance and documented in a permanent record in Prospero (Registration number CRD42018115828).
Cohort studies, case control studies, case reports and case series were included. No publication date or publication status restrictions were imposed. Patients of any age were included if a diagnosis of rhabdomyolysis was made and in addition the diagnosis of NMS or SS was considered definite or probable. A separate search was performed on all cases of rhabdomyolysis induced by the use of the recreational drug 3,4-methylenedioxymethamphetamine (MDMA/ecstasy), since not all literature considers this as a specific form of SS [14] . In cases where CK levels were unknown, or serum CK levels did not exceed 1.5 times the upper limit of normal (ULN), the study was excluded as not fulfilling the diagnostic criteria for hyperCKemia ( > 1.5x ULN) or rhabdomyolysis ( > 10x ULN) [15] . Meta-analyses, review articles, conference abstracts, articles not in English, animal studies, and studies of which no full-text was available, were also excluded.
A literature search was performed in MEDLINE, EMBASE, Psycinfo and the Cochrane Library from inception to January 31, 2019. The following keywords and all known synonyms for these keywords were used: 'Neuroleptic malignant syndrome', 'Serotonin syndrome', 'MDMA', 'Rhabdomyolysis'. In addition, based on clinical similarities and in particular considering rhabdomyolysis as a shared clinical endpoint, we conducted an additional search for RYR1 , given that the part of the NMS-and SS-related rhabdomyolysis spectrum is possibly associated with RYR1 variants. The search strategy is detailed in Appendix 1. Eligibility assessment was performed by two independent reviewers (N.K, J.W) in an unblinded standardized manner. Each study was reviewed for eligibility based on title and abstract. The full text of presumably eligible studies was evaluated to decide whether the study fulfilled the inclusion criteria. In case of discrepancies between the two reviewers, consensus was reached after discussion; if no agreement could be reached, a third reviewer was consulted. We developed a data extraction sheet, which was then pilot-tested on ten randomly-selected included studies, and subsequently refined accordingly. One review author (N.K.) extracted the data items from included studies. Information was extracted from each included study on (1) clinical characteristics (including age, sex, medical history, family history, temperature and mental status during the episode); (2) laboratory results (including peak CK levels, myoglobinuria, genetic testing); (3) neurological symptoms and examination (including myalgia, muscle cramps/weakness/swelling, extrapyramidal symptoms, hyperreflexia, clonus); (4) outcome (including recurrence, death). The number of episodes was divided into three groups: recurrent, non-recurrent and unknown. If multiple episodes of rhabdomyolysis were reported in one patient, the first episode was used for describing the data.
Based on our experience with exertional rhabdomyolysis, the following clinical features were considered suggestive for a monogenic background: 1) recurrent rhabdomyolysis and/or 2) a positive family history of neuromuscular disorders, in particular MH and rhabdomyolysis [16] .
Categorical data were compared using the χ 2 test, or Fisher exact test in case of less than 5 patients. Continuous variables were compared using the Mann-Whitney U test.

Results
The search yielded a total of 1206 studies. After removal of duplicates, 976 studies remained. After reviewing the titles and abstracts, 699 studies were excluded because they did not meet the eligibility criteria. The full text of the remaining 277 studies was examined in more detail. It appeared that 168 citations did not meet the inclusion criteria ( Fig. 1 ).

Study characteristics
A total of 109 studies (10 case series and 99 case reports) were included, comprising 143 cases of which 134 cases met our inclusion criteria. These included 60 cases of SS-and 74 cases of NMS-related rhabdomyolysis.

Clinical features
Characteristics of the patients are described in Table 1 . The median age was 35.0 years (range, 4-84). 77 (58%) were male, 57 (43%) were female. Death was reported in 23 patients (17%). In eight patients, rhabdomyolysis was recurrent (6%). Myoglobinuria was reported in 38 patients (27%). Results of the neurological examination during the episode of rhabdomyolysis are listed in Table 2 .

Serotonin syndrome
A total of sixty patients with SS, 33 male and 27 female, were included. Those patients had a median age of 23 years (range 4-75). The median maximum temperature during the episode was 39.5 °C (range 35.4-43.3). The median peak CK level was 10,888 (range 354-196,000 IU/l). Thirty three cases could be attributed to the prescription of a specific medication, twenty seven cases (45%) involved SS due to intoxication with MDMA. Patients with MDMA intoxication were younger than patients in whom medication was

Neuroleptic malignant syndrome
In the group of 74 patients with NMS, 44 were male and 30 were female; they had a median age of 45 years (range 9-84). Patients had a median maximum temperature of 39.5 °C (range 35.4-42.8). The median peak CK level was 15,059 IU/L (range 572-220,000). Rigidity was reported more  Table 3 shows the characteristics of the subset of eight patients experiencing recurrent episodes of NMS and/or SS. Table 4 shows the characteristics of the individual cases.

Presumed genetic susceptibility and genetic screening
Genetic testing was performed in one case report and one case series, comprising eleven patients (8%).

Serotonin syndrome
Out of 60 patients with SS, two patients developed recurrent episodes of rhabdomyolysis. Both patients developed SS and also had experienced an episode of NMS in their history or at follow up (patients 1 and 2, Table 4 ). Of these, genetic screening was performed in one patient. This patient, a 20-year-old woman, developed rhabdomyolysis after exposure to MDMA [17] . Two weeks later, she developed NMS after being exposed to haloperidol. Performing both RT-PCR, covering the complete RYR1 transcript, and sequence analysis of genomic DNA, a presumed pathogenic variant in the RYR1 gene was revealed (c.2545G > A in exon 20 resulting in an amino acid change p.Asp849Asn). Her asymptomatic mother carried the same variant and was considered MH negative after caffeinehalothane contracture testing.

Neuroleptic malignant syndrome
Recurrent episodes of rhabdomyolysis due to NMS were reported in six patients (patients 3-8, Table 4 ). Seven patients had experienced an adverse drug reaction (other than rhabdomyolysis) to antipsychotics in their history. One case series included limited RYR1 genetic testing in ten patients with NMS. Six RYR1 variants known to be associated with malignant hyperthermia were analyzed, with negative test results. However, comprehensive RYR1 screening was not performed [10] .
In 129 cases (96%) no information regarding family history was provided, and five patients mentioned a negative family history. In one patient, a debatable positive family history involved an asymptomatic mother carrying a RYR1 variant (patient 1, Table 4 ).

Discussion
This systematic review identified 109 studies fulfilling the selection criteria, comprising 74 cases of NMS and 60 cases SS associated with hyperCKemia or rhabdomyolysis.
In both subsets, a male predominance of 60% (NMS) and 55% (SS) was found. Cases with SS more often involved an auto-intoxication with medication (20%) or recreational drugs (45%) than cases with NMS (4%) ( p < 0.001), reflecting that SS is a concentration-dependent syndrome and NMS is considered an idiosyncratic reaction [18] . Eight patients experienced recurrent episodes of rhabdomyolysis. Those involved six patients with recurrent NMS, and two patients with SS that had experienced NMS at different timepoints. Many of the clinical characteristics of both syndromes overlap. It has been discovered that the pathways of serotonin and dopamine are very much intertwined, and it may be therefore that symptoms associated with both SS and NMS are reflected by a final shared pathway [19] . However, hyperreflexia and convulsions were reported more often in patients with SS, both features supportive of central origin. A summary of the clinical phenomenology of both syndromes is depicted in Fig. 2 .
The mean age in our subset of patients with NMS is similar to the average median age of 39 years found in a recent systematic review involving also NMS patients without rhabdomyolysis [20] . The age of patients with SS in our review is below the mean age of 58 found in a large French retrospective study, also after separating MDMA-induced SS from medication induced SS (mean ages of 23 and 35, respectively) [21] . For both NMS and SS, the mean age was higher compared to the age of onset in patients with MH, experiencing a first adverse reaction at a mean age of 18 years [22] . Presumably, this is due to the prescription of antipsychotics and antidepressants at an older age, or the prescription of multiple neuroleptics at a higher age, compared to potentially MH-triggering anaesthetics and muscle relaxants that are administered throughout lifetime, or even preferentially in the pediatric age group.
The male preponderance seen in our review is also observed in malignant hyperthermia, not in keeping with what is expected in a strictly monogenic autosomal-dominant disorder [ 23 , 24 ]. It has been previously hypothesized that the male predominance in MH was caused by men being more frequently exposed to types of surgery commonly associated with MH. However, a study on in vitro contracture test (IVCT) outcome reported a significant difference in sex distribution, with males revealing a positive IVCT more frequently [24] , probably reflecting additional hormonal or other sex-dependent variables. In addition, there is a higher use of antipsychotics in men compared to women, and men are known also to adhere to medication more consistently [25] . Also, in a study on acute intoxications in six Dutch hospitals, intoxications due to substance abuse was more frequently seen in men (66.0%) [26] .
The mortality rate in both subsets of patients was 20% (NMS) and 13% (SS), respectively. Mortality rates in MH have decreased from 80% thirty years ago, to < 5% in 2006 [23] , because of the introduction of Dantrolene into pharmacological MH management. In more recent studies in our cohort, less deaths were reported in both groups (mortality of 13% from 2000 to 2018, 27% from 1984 to 1999).
The cumulative recurrence rate of 6% in our present study (8% in NMS and 3% in SS) is lower than in studies on rhabdomyolysis in general. In a retrospective study performed in 2005, 475 hospitalized patients with rhabdomyolysis were described, reporting a recurrence rate of 11% in a group with various etiologies [27] . A subgroup analysis of 38 NMS patients in this study revealed a recurrence rate of 25%. No investigations into possible underlying genetic susceptibilities were performed in these subjects. In 56% of all cases in our study, information on possible previous episodes in the patient's medical history was lacking, likely pointing to an underestimation of recurrence rates. In patients with recurrent Table 5 Overview of genetic testing in patients with episodes of rhabdomyolysis due to psychiatric agents in our present study, and previous studies. episodes, no discernable factors (for example, presence of intercurrent illness, fever, exercise) were reported that may have modified the presentation. The mother of patient 1 carried the same RYR1 variant (c.2545G > A in exon 20 resulting in an amino acid change p.Asp849Asn) but was tested negative for caffeine-halothane contracture testing (CHCT), leading the authors to conclude that the RYR1 variant was non-contributory to the phenotype. However, a recent study shows that the majority of patients with ERM related to pathogenic RYR 1 variants had a negative CHCT (or in vitro contracture test (IVCT)), indicating that these tests do not have the same discriminatory value for RYR1 -related rhabdomyolysis as for RYR1 -related MH [28] . Interestingly, a case report published in 2003 reported a patient with NMS in whom contracture testing with caffeine and halothane was negative, but in an analogous setting, the patients isolated muscle was found to react pathologically towards contracture testing with neuroleptics [29] .
We also identified additional cases in the literature in whom rhabdomyolysis was triggered by exposure to antipsychotics, but who were not diagnosed as NMS or SS and were therefore not included in the current review. These patients are summarized in Table 5 . Dlamini et al. sequenced RYR1 in 39 unrelated families with rhabdomyolysis and identified 9 heterozygous variants in 14 families [2] . Two patients were on olanzapine treatment at the time of presentation with rhabdomyolysis. In both patients, a RYR1 variant was found, one of which the authors mentioned to be associated with MH (p.Tyr2426Cys). However, to date, the relation of this variant with MH remains unclear, although it localizes to the MH mutational hotspot affecting the central domain of the RyR1 protein. The second patient on olanzapine treatment involved the p.Thr4288_Ala4290dup variant, that has been reported in three MHS individuals of which two also had a history of exercise induced rhabdomyolysis. Since all these individuals also had a second RYR1 variant, the specific role of the p.Thr4288_Ala4290dup in MH and rhabdomyolysis remains uncertain [30] .
In 2010, Sato et al. performed post mortem genetic analysis on eleven psychiatric patients who died of hyperthermia, most likely caused by NMS [31] . This paper was not included in our quantitative data-analysis because it did not report CK levels. Two patients carried a RYR1 variant resulting in an amino acid change: 1) p.Arg4645Gln, a variant probably not pathogenic, based on evolutionary conservation, and 2) p.Ala612Thr, a substitution affecting a conserved amino acid residue that might have plausibly contributed to the event of rhabdomyolysis, regarding its location near the common p.Arg614Cys diagnostic MH variant.
To our knowledge, this is the first literature review describing all reported patients with hyperCKemia or rhabdomyolysis in the context of NMS and SS. However, the quality of data obtained from most reports was considered low; neurological examination was often not or only incompletely described, resulting in a substantial amount of missing data. In addition, information regarding recurrent episodes family history, or genetic testing was often absent. Therefore, it is impossible to ascertain if or if not some clinical characteristics were not present, or if the author decided not to describe a negative test result. Furthermore, only few studies included information concerning the patient's medical and family history.
Genetic testing was performed in only a small amount of studies of patients (8%). Therefore, based on the currently available data, there is little evidence to suggest that all patients should undergo genetic testing, however, these recommendations may change if additional information based on more comprehensive genetic testing in such cohorts becomes available. Although highly desirable, considering the comparative rarity of these presentations, performing a prospective study is not feasible. Based on the results in our present study we would advise to perform additional genetic testing, for example through an NGS-based rhabdomyolysis panel, in patients with drug-induced rhabdomyolysis and features suggestive of genetic susceptibility, such as a positive family history or recurrent episodes. To inform such an approach, it is, firstly, important to recognize specific genotype-phenotype correlations of genetic disorders associated with rhabdomyolysis to further determine the diagnostic approach, and to decide which candidate genes are opportune for testing [32] . Secondly, although the RYR1 gene is a likely candidate for these presentations, RYR1 is also an extraordinarily complex gene with over 5000 amino acid residues that harbors many, often common missense variants (over 2000 already known) [33] . Performing next generation sequencing on a large gene such as RYR1 will therefore in all likelihood reveal a large number of variants of unknown significance. Although RYR1 analysis therefore requires a very careful and informed approach, in our view RYR1 sequencing should be considered in particular in patients with recurrent episodes, considering the presence of RYR1 variants in a proportion of the patients with NMS-or SS-related rhabdomyolysis we found in our literature review.
Identification of the genetic background implicated in these pharmacogenetic syndromes is important for several reasons. Firstly, in line with a personalized medicine approach, an informed decision can be made to avoid certain pharmacological agents in individuals where those may be harmful due to a specific genetic predisposition. Secondly, specific measures may be taken (for example, prescription of the specific ryanodine receptor antagonist Dantrolene in cases where RYR1 has been implicated) to prevent recurrent episodes [34] . Thirdly, in cases where diagnostic MH variants in RYR1 have been identified, genetic counseling of the patient and their families may prevent life-threatening MH events. The potential to identify the genetic background of NMS and SS has massively increased in recent years, due to the increasing availability of next generation sequencing, but appears not to have been fully utilized yet in these conditions.

Conclusion
Many of the clinical characteristics of SS and NMS overlap, except for hyperreflexia and convulsions (SS) and rigidity (NMS). This review underlines that a subset of patients develop recurrent episodes of rhabdomyolysis during NMS or SS, or both syndromes during different timepoints, suggesting a possible underlying (genetic) susceptibility. The genetic background of NMS and SS has been investigated only to a very limited degree so far, but the increasing availability of next generation sequencing offers an opportunity to identify potentially associated genetic backgrounds, especially in patients with a positive family history or recurrent episodes.