Systemic corticosteroids for the treatment of acute episodes of rhabdomyolysis in lipin‐1‐deficient patients

Mutations in the LPIN1 gene constitute a major cause of severe rhabdomyolysis (RM). The TLR9 activation prompted us to treat patients with corticosteroids in acute conditions. In patients with LPIN1 mutations, RM and at‐risk situations that can trigger RM have been treated in a uniform manner. Since 2015, these patients have also received intravenous corticosteroids. We retrospectively compared data on hospital stays by corticosteroid‐treated patients vs. patients not treated with corticosteroids. Nineteen patients were hospitalized. The median number of admissions per patient was 21 overall and did not differ when comparing the 10 corticosteroid‐treated patients with the 9 patients not treated with corticosteroids. Four patients in the non‐corticosteroid group died during a RM (mean age at death: 5.6 years). There were no deaths in the corticosteroid group. The two groups did not differ significantly in the number of RM episodes. However, for the six patients who had RM and occasionally been treated with corticosteroids, the median number of RM episodes was significantly lower when intravenous steroids had been administered. The peak plasma creatine kinase level and the area under the curve were or tended to be higher in patients treated with corticosteroids—even after the exclusion of deceased patients or focusing on the period after 2015. The median length of stay (10 days overall) was significantly longer for corticosteroid‐treated patients but was similar after the exclusion of deceased patients. The absence of deaths and the higher severity of RM observed among corticosteroid‐treated patients could suggest that corticotherapy is associated with greater survival.


Funding information
Agence Nationale de la Recherche, Grant/Award Number: ANR-13-BSV1-0020-01; Association Française contre les Myopathies, Grant/Award Numbers: 15947, 13988; Fondation Maladies Rares, Grant/Award Number: 09071 Communicating Editor: Areeg El-Gharbawy situations that can trigger RM have been treated in a uniform manner. Since 2015, these patients have also received intravenous corticosteroids. We retrospectively compared data on hospital stays by corticosteroid-treated patients vs. patients not treated with corticosteroids. Nineteen patients were hospitalized. The median number of admissions per patient was 21 overall and did not differ when comparing the 10 corticosteroid-treated patients with the 9 patients not treated with corticosteroids. Four patients in the non-corticosteroid group died during a RM (mean age at death: 5.6 years). There were no deaths in the corticosteroid group. The two groups did not differ significantly in the number of RM episodes. However, for the six patients who had RM and occasionally been treated with corticosteroids, the median number of RM episodes was significantly lower when intravenous steroids had been administered. The peak plasma creatine kinase level and the area under the curve were or tended to be higher in patients treated with corticosteroids-even after the exclusion of deceased patients or focusing on the period after 2015. The median length of stay (10 days overall) was significantly longer for corticosteroid-treated patients but was similar after the exclusion of deceased patients. The absence of deaths and the higher severity of RM observed among corticosteroid-treated patients could suggest that corticotherapy is associated with greater survival.

K E Y W O R D S
corticosteroid, lipin-1 deficiency, rhabdomyolysis, survival

| INTRODUCTION
Lipin-1 deficiency (caused by mutations in the LPIN1 gene) is a rare, life-threatening condition that causes episodes of rhabdomyolysis (RM). Whereas severe RM is typically defined by a plasma creatine kinase (CK) level >6000 IU/ L, the level in lipin-1 deficiency is usually >50 000 IU/L. RM is typically triggered by fever, fasting, and (more rarely) effort. 1,2 LPIN1 mutations are associated with a poor prognosis, despite aggressive management in the intensive care unit (ICU). The complications of RM include cardiac arrhythmia (due to hyperkalaemia) and kidney failure. [2][3][4][5] The cardiac complications are possibly linked to lipin-1's specific regulation of the heart's lipid metabolism and contractile function in response to hemodynamic stress. [6][7][8] Accordingly, patients with LPIN1 mutations accumulate neutral lipid in the myocardium. 6 Cardiac lipotoxicity has been implicated in the development of cardiac dysfunction in a variety of models. 8 Disease onset occurs usually in early infancy and rarely in adulthood. 9 Although episodes of RM are more frequent during infancy (when the incidence of triggering viral infections is highest), the risk remains into adulthood. The incidence of RM episodes per patient ranges from one to several dozen. CK levels are normal or above-normal between RM episodes.
The only available treatment for RM is best supportive care in the ICU, with aggressive intravenous hydration and monitoring for complications; these approaches can reduce the duration of RM episodes from 7-10 days to 5 days. 10 Since RM episodes are triggered by febrile illnesses (conditions that feature high circulating levels of proinflammatory mediators), we suspected that inflammation has a major role in lipin-1 deficiency. 11 Lipin-1 is abundantly expressed in adipocytes and skeletal muscle. It acts as both (i) a phosphatidate acid phosphatase (PAP1) that dephosphorylates phosphatidic acid to diacylglycerol and thus contributes to triacylglycerol and phospholipid biosynthesis, 12 and (ii) a transcriptional co-activator of PPARα, SREBP1, and PGC-1α (all of which are involved in energy pathways). 13 Coactivator expression patterns were normal in our patients' adipocytes 14 and skeletal muscle. 11 We recently showed that the pathogenic sequence linking lipin-1 deficiency and RM results in the loss of PAP1's enzymatic activity, which in turn leads to a reduction in phosphatidylinositol 3-kinase-Vps34 and phosphatidylinositol 3-phosphate (PI3P) levels specifically at the membrane of late endosomes in the patients' cells. 15 The loss of PI3P alters mitophagy in animal models 16 and in human myoblasts via an abnormal GTP-bound Rab7, 15 resulting in the accumulation of mitochondrial DNA (mtDNA), a natural ligand of the Toll-like Receptor 9 (TLR9) 17 in late endosomal/lysosomal structures of patient myoblasts exposed to metabolic stress. In turn, this results in TLR9 activation, elevated secretion of inflammatory cytokines, the release of calcium into the cytoplasm, and cell death; the latter two features are hallmarks of RM. 15 As a new inflammatory disease, lipin-1 deficiency reinforces the notion of immunometabolism-an emerging interface between inflammation and cell metabolism. 18 This notion prompted us to treat our patients' acute episodes of RM with corticosteroids as already proposed in other RM. 19 Our hypothesis was confirmed by (i) the presence of proinflammatory CD4+ and CD8+ lymphocytes within the muscle tissue of one patient, 5 (ii) the correction of lipid accumulation after the corticosteroid treatment of patient myoblasts cultured with proinflammatory cytokines, 11 and (iii) the beneficial role of dexamethasone in a few lipin-1 deficient patients with acute RM. 5,15,20,21 In order to demonstrate the clinical benefit of corticosteroids in the management of patients with lipin-1 deficiency, we retrospectively collected data from hospital admissions for RM and/or triggering factors (fever, gastroenteritis, etc.). We compared clinical courses, plasma CK levels, and the length of the hospital stay for lipin-1 patients treated with corticosteroids vs. those not treated with corticosteroids.

| PATIENTS AND METHODS
We collected medical data from French hospitals where lipin-1-deficient patients were admitted in emergency units (non-scheduled admission) for an episode of RM or an at-risk situation (i.e., "preventive" hospital stays with a RM risk factor such as fever, viral illness, fasting, etc., regardless of the CK level). Hereafter, we shall refer to these non-scheduled admissions simply as "admissions".
In all cases, the diagnosis of lipin-1 deficiency was confirmed by a PCR analysis of DNA isolated from blood (as described previously 2 ) or by next-generation sequencing with an HiSeq system (Illumina) and Agilent Technologies probes. Moderate RM and severe RM were defined as a plasma CK level > 1000 IU/L and >6000 IU/ L, respectively (value in healthy controls: <160 IU/L).
These emergency situations (i.e. an acute episode of RM or an RM risk factor) were managed in the same way at all participating hospitals, thanks to the use of a common, standardized protocol which was common to all RM whatever their cause. In view of the seriousness of the disease, most patients were admitted to an emergency unit or an ICU for at least 24 h (when CK levels were normal) or 48 h (if patients with normal CK levels experienced muscle pain). Patients were continuously monitored throughout the hospital admission. Hydration in cases of RM (abnormal CK levels) included saline fluid resuscitation (20 mL per kg) and then 3 L/m 2 /day [body surface area = (4 Â W + 7)/(W + 90)] without waiting for an increase in potassium or CK. One liter of solute consisted of 200 mL of 30% glucose (180 g/m 2 , 6 g/ kg/24 h), 400 mL of 1.4% isotonic bicarbonate, and 400 mL of 0.9% NaCl, with no potassium or calcium. Hemodialysis was considered in cases with hyperkalaemia (≥5 mmoL/L) despite effective hyperhydration, electrocardiogram (ECG) abnormalities, anuria or oliguria (<0.5 mL per kg/h), or symptomatic, lifethreatening fluid overload (with an overweight of >20%) contraindicating the continuation of hyperhydration. If changes in the ECG suggesting hyperkalaemia, the patient was given intravenous 1.4% isotonic bicarbonate (10 mL/kg [1.5 mmol/kg]), intravenous calcium with 0.1 mmol/kg of elemental calcium; and for patients with persistent, life-threatening ECG abnormalities: 30 mL/kg of 14% isotonic bicarbonate per cannula for intraperitoneal administration, prior to the initiation of continuous renal replacement therapy as soon as possible. In the absence of RM (normal CK levels), only 1.5-2.5 L/m 2 /day of the same solution were administered.
We monitored the following parameters following the patient's admission to hospital: plasma CK and blood ions (Na and K) every 4 to 12 h for the first 24 h, glycemia, a complete blood ion assessment with Ca, Ph, Mg, urea, and creatinine (every 6 h); hourly urinary output >2 mL/kg/h; fluid balance assessment every 3 h (in order to adjust treatment, if required), arterial blood pressure, regular ECG, and at least one echocardiography.
The criterion for discharge from hospital were the same at all participating hospitals: the absence of CK elevation after 24-48 h, or a return to normal or subnormal plasma CK levels (<2000 IU/L) after RM.
Since 2015, most patients hospitalized for an episode RM or a risk factor for RM have been treated with corticosteroids immediately upon admission as a routine care through the modification of their emergency certificate. As soon as the signs and symptoms of acute RM or infections (muscle pain, and/or fatigue, fever, gastroenteritis that can trigger RM) appeared, patients typically took an initial oral dose of prednisone (Cortancyl ® [Sanofi Aventis, Gentilly, France], 1-2 mg/kg) at home (after phone authorization by a physician) before going to hospital as soon as possible. Intravenous treatment with a corticosteroid (methylprednisolone) was immediately initiated with a single dose of 1-2 mg/kg/day for the first 3 days. However, one patient received a single dose of 2 mg/kg/day for 4 days and then a single dose of 4 mg/kg/day from day 5 to day 20 due to the severity of the episode and its duration. Any adverse reactions to corticosteroids (such as susceptibility to infections, metabolic and endocrine effects [hypokalaemia, and pronounced hyperglycaemia needing insulin therapy], and severe hypertension requiring treatment) were noted.

| Statistical analyses
For each hospitalization, we calculated the length of the hospital stay and examined the distribution of plasma CK levels measured during the stay (i.e. the peak CK value and the area under the curve). We also summarized these parameters at the patient level. Lastly, we quantified the numbers of episodes of moderate RM and severe RM for each patient.
Categorical variables were expressed as the frequency (percentage). Continuous variables were expressed as the means (standard deviation [SD]) or the median [interquartile range (IQR)] and range. Fisher's exact test or the Mann-Whitney test were used to compare survival status, length of the hospital stay, and CK parameters in patients having received intravenous corticosteroids vs. patients not having received intravenous corticosteroids. For the patients who had received corticosteroids during some (but not all) admissions, Wilcoxon's signed-rank test was applied.
As sensitivity analyses, we excluded deceased patients, and we performed the analyses on the period after 2015.
All statistical analyses were performed with STATA software (release 15.0; Stata Corporation, College Station, TX, USA).

| Patients' characteristics
Nineteen patients with LPIN1 mutations (born between 1995 and 2016; 10 males and 9 females from 13 different families, including three sets of three siblings) were admitted in emergency unit (non-scheduled admission). The characteristics of the admissions are summarized in Table 1 (the number of hospitalizations and RM, length of the hospital stay, and plasma CK levels) and Table S1 (variables related to RM). The age at the time of the first ever episode of RM ranged from 1.5 to 19.4 years, and the mean (SD) value was 5.6 (4.8). The median (IQR) peak CK level was 210 800 (450014) IU/L. Some patients had a history of myalgia, muscle cramps, fatigability, exercise intolerance, and (more rarely) cardiomyopathy. 6 Nine patients did not receive any corticosteroids during any of their admissions, seven patients received corticosteroids during some (but not all) admissions (including six patients who had at least one RM treated with corticosteroids and one patient treated with corticosteroids during

| Patient survival
Four patients (P16, P15, P17, and P7; two males and two females) died at hospital in 2000, 2010, 2011, and 2016, respectively, as a result of severe RM 3 ; the mean (SD) (range) age at death was 5.6 (0.7) (4.9-6.3) years. None of the four patients had been treated with methylprednisolone ( Table 1). The diagnosis of lipin-1 deficiency was known for three patients (P16, P17, and P7) for whom the standardized protocol was strictly applied especially with immediate hospitalization in case of triggering event, in view of disease severity. However, this strict management did not prevent the death of these patients. P16 who died in 2016 was unfortunately treated without corticosteroids. The patient was well known to his university hospital, and the usual emergency protocol was applied but without being reviewed. The emergency workers also did not call our on-call telephone number, therefore we could not insist on the need to administer corticosteroids. P15 died at his first RM, 3 the diagnosis of lipin-1 deficiency being unknown before.

| Hospitalization characteristics
Median (IQR) number of admissions per patient was 21 (14); there was no difference between patients having received corticosteroids and those who not having received corticosteroids (p = 0.346) ( Table 1). There were no intergroup differences in the number of admissions with moderate RM (CK > 1000 IU/L), with severe RM (CK > 6000 IU/L) and without RM. Similar results were T A B L E 2 Characteristics (related to RM, plasma creatine kinase levels, and the hospital stay) of 138 admissions by the six patients with LPIN1 mutations and RM who had been treated occasionally with corticosteroids.  [1] in the absence of corticosteroids; p = 0.033, Wilcoxon's signed-rank test). The peak plasma CK level and the area under the curve were higher in the group of patients who had received corticosteroids, although this difference was not significantly significant (p = 0.191 and p = 0.051, respectively). The same was true when the deceased patients were excluded from the analysis (p = 0.178 and p = 0.086, Table 1).
The median (IQR) length of the hospital stay was 10 (10.3) days overall and was significantly longer for patients treated with corticosteroids than those treated without corticosteroids (p = 0.041), mainly due to the short length of the hospital stay for deceased patients. After the exclusion of deceased patients from the analysis, the median (IQR) length of the hospital stay fell to 11.9 (8.1) days and there was no difference between the corticosteroid and non-corticosteroid groups (p = 0.327).
When focusing on the period after 2015, no intergroup difference in the number of admissions with RM and without RM was observed. However, the peak plasma CK level and the area under the curve were higher (tendency or a significant difference) in patients who received corticosteroids. Furthermore, the length of the hospital stay for episodes of severe RM was significantly longer when corticosteroids were administered ( Table 1).
The data per admission are summarized in Table 3. The 19 patients had a total of 313 documented admissions with 84 episodes of RM (including 30 with moderate RM and 54 with severe RM). There were 290 admissions without corticosteroid treatment and 23 admissions with corticosteroid treatment. The severity of RM was linked to the use of intravenous steroid therapy because the proportion of admissions with severe RM (as judged by the peak CK levels and the area under the CK curve) was higher when corticosteroids were administered. Furthermore, the length of the hospital stay for episodes of severe RM was significantly longer when corticosteroids were administered, although this relationship was no longer significant when deceased patients were excluded from the analysis (6.0 Accordingly, to the severity of RM treated by corticosteroids, we could note a compartment syndrome involving both lower limbs, requiring fasciotomy with hemodialysis and intubation in one episode; short ventricular extrasystoles during the placement of a central venous access port in another one. This was followed by hyperkalaemia, acute kidney failure, systemic hypertension and posterior reversible encephalopathy syndrome, necessitating hemodialysis, intubation, and vasopressor treatment over 2 days; neurological criteria prompted intubation for another one. No adverse drug reactions related to corticosteroids (infections, severe hypertension requiring treatment, hypokalaemia, or pronounced hyperglycaemia requiring insulin therapy) were noted.
3.4 | Focus on seven patients with severe rhabdomyolysis that led to death or were cured by corticosteroid treatment Patient 7 was the eldest of the siblings P7-P8-P9 and passed away at the age of 6 years in 2016, following a severe episode of RM that was not treated with corticosteroids. He had previously experienced severe episodes of RM at the ages of 2, 4, and 6 years (peak CK level at the age of 2 years: 193 300 IU/L). During the fourth episode of RM (triggered by acute gastroenteritis), the patient developed a slight fever; this prompted hospital admission and hyperhydration. Salbutamol was administered , neurological criteria prompted intubation. Kalaemia was below 5.0 mmol/L. Acute kidney failure did not occur. Patient 19 was hospitalized three times at the age of 3.5 years and experienced two episodes of severe RM (peak CPK level = 36 824 and 447 300 IU/L, respectively). The second episode of RM was triggered by a pneumococcal infection and required a 21-day stay in the ICU. Patient 19 experienced short ventricular extrasystoles during the placement of a central venous access port. This was followed by hyperkalaemia (6.1 mmol/L), acute kidney failure, systemic hypertension, and posterior reversible encephalopathy syndrome, necessitating hemodialysis, intubation, and vasopressor treatment over 2 days. Methylprednisolone was administered over 3 weeks at an initial dose of 4 mg/kg/day and then 2 mg/kg/day. Patient 19 received prednisone and methylprednisolone during his third hospital stay and did not develop RM.
Patient 17 experienced two severe episodes of RM at the ages of 2 and 2.5 years (peak CPK level = 690 760 and 255 200 IU/L, respectively), which were complicated by systemic hypertension and kidney failure. During the third episode of RM, the patient was admitted to the ICU after experiencing diffuse muscle pain in the morning. The ECG showed diffuse, symmetrical, high-amplitude T waves. The blood potassium level was 5 mmol/L, and arrythmia soon developed. The peak CPK level was 57 830 IU/L. Despite the application of an intensive resuscitation protocol, the patient died within hours of ICU entry. Corticosteroids were not administered.
At the age of 6 years, P15 3 presented at the ICU with diffuse muscle pain and stiffness but stable respiratory and hemodynamic vital signs. The peak CPK level was 22 013 IU/L, and his last kalaemia before arrythmia was 4.8 mmol/L. The patient suddenly developed ventricular tachycardia and then cardiorespiratory arrest. Despite the application of an intensive cardiopulmonary resuscitation protocol, the patient died a few hours later.
Patient 16 experienced five severe episodes of RM between the ages of 2 and 4 years, with a peak CPK level of 210 800 IU/L. During the fifth episode, the girl (aged 5) died following cardiorespiratory arrest, with a peak CPK level of 25 020 IU/L, and subnormal kalaemia.

| DISCUSSION
Despite our intensive treatment of episodes of acute RM or at-risk situations, 3 the mortality rate among patients with lipin-1 deficiency was high (as emphasized by P7's death in 2016). Most deaths were due to arrhythmia, [3][4][5] which was probably related to hyperkalaemia and/or cardiac dysfunction. Indeed, lipin-1 has specific roles in cardiac lipid metabolism. [6][7][8] Hence, there was an urgent need for a treatment that prevented or attenuated RM and its complications. Based on our findings for lipin-1 deficiency (a pro-inflammatory process with TLR9 hyperactivation due to a phospholipid defect and a blockage in mitophagy), and experience for other RM, 19,20 we proposed short course of systemic corticosteroids following admission for acute situations and we retrospectively evaluated their effectiveness and safety. Corticosteroids decrease capillary permeability and the transcription of inflammatory mediators and therefore reduce the overall inflammatory load. 19 The treatment consisted of the intravenous administration of methylprednisolone (at a dose of 1 to 2 mg/kg/day or, in one case, 4 mg/kg/day) in a hospital setting, together with intensive care. Most of the treated patients also received oral prednisone at home before admission to hospital; this might have saved time, notably because complications can occur soon after the onset of RM. 3 Remarkably, none of the corticosteroid-treated patients died; all the deaths resulted from acute episodes of RM in patients not treated with corticosteroids (including three deaths since 2010 and one in 2016; the diagnosis was known for three patients). Moreover, when considering the six patients who had been treated occasionally with corticosteroids, the median number of episodes of RM was significantly lower when corticosteroids were administered orally (before admission to hospital) and then intravenously (in hospital). Even though the episodes of RM were more frequently severe (according to the peak plasma CK level and the area under the curve) in corticosteroid-treated patients, the median length of the hospital stay for patients with moderate RM and those with severe RM were similar after deceased patients had been excluded from the analysis. These differences were also obtained when focusing on the period after 2015. These data (i.e. no death and a similar length of the hospital stay for more severe RM) could suggest that corticosteroids (together with aggressive organ support and the prevention of at-risk situations, i.e. hospitalization in all cases) help to effectively improve clinical outcomes. These episodes of MR might have led to death without corticosteroids. We have practised "preventive" hospitalizations for a long time in the field of inherited metabolic diseases and we believe it is still an appropriate method in diseases at risk of RM such as mitochondrial fatty acid disorders. In the case of lipin-1 deficiency, even before the LPIN1 molecular diagnosis, the medical staff and the parents were rapidly aware of those RM episodes seriousness therefore this preventive protocol has been rigorously applied for a long time. Thus, our hospital admission protocol has changed little over time, except for the administration of corticosteroids and (perhaps) stricter monitoring.
It should be noted that the potassium level is probably not a good marker of severity because it was sometimes subnormal or moderately elevated before the death of a patient, and not necessarily correlated with the ECG pattern 5 (see focus on seven patients).
Our study has several limits: it is a retrospective and not controlled study of a small sample size (not a prospective randomized well-designed comparative-controlled study) and the period without corticosteroids was longer than the period with corticosteroids administration. While treatments during admissions were standardized, we cannot affirm that there was no variation in the management with time. In the same way, although RM treated with corticosteroids were more severe than those not treated with corticosteroids, we cannot exclude that the number of deaths was not due to the larger group (admissions) not treated with corticosteroids. However, when focusing on the period after 2015, the peak plasma CK level and the area under the curve were also higher (tendency or a significant difference) in the group of patients who had received corticosteroids. Furthermore, the length of the hospital stay for episodes of severe RM was significantly longer when corticosteroids were administered.
Given the good safety profile observed during the 23 admissions with corticosteroid administration (in 10 patients), we have continued to give corticosteroids at home and then upon admission to hospital in all atrisk situations. The treatment is simple, inexpensive, and easy to access. The speed of death of patient P15 also shows the interest of corticosteroids at home just before being hospitalized, to save time. It is a real therapeutic education. Given that corticosteroids have been linked to adverse events like susceptibility to infections and metabolic and endocrine effects (albeit not observed in our cohort), the dose and duration of corticosteroid treatment must be determined on a rigorous basis (i.e. taking account of contraindications, infections, and the riskbenefit ratio) during an admission that features other therapeutic and preventive measures. We recommend vaccinating lipin-1-deficient patients with all available vaccines, as the risk of decompensation is very high during epidemics of viral disease, and corticosteroids can make chickenpox worse. We include the vaccination against COVID-19, 22 although PAP-1 deficiency significantly suppresses SARS-CoV-2 replication 23 and this vaccine can be responsive for RM. 24 The innate immune system acts as the first line of defense against invading pathogens. Induction of an antiviral innate immune response depends on the TLRs, which then activate signaling pathways and induce the production of antiviral cytokines and chemokines. TLRs are present at the cell surface or in endosomal compartments; for example, TLR9 that is involved in the recognition of single-stranded DNA (hypomethylated CpG motifs, specifically) and then upregulates inflammatory mediators. 25 When the mitochondria are stressed or damaged, these organelles generate stress signals that include ROS and mtDNA release. Cell-free mtDNA acts as a damage-associated molecular pattern (due to its similarity to bacterial ancestors) and activates several innate immune pathways (including the TLR9 pathway). 17 Mitophagy that removes damaged mitochondria and thus is crucial for preserving mitochondrial and immune homeostasis. When mitochondria are defective, for example, in lipin-1 deficiency, they trigger inflammatory cascades. Interestingly, the fact that gastrointestinal and respiratory viruses are frequently found in cases of LPIN1-mediated RM emphasizes the role of viral triggers via TLR9. 26 It should be noted that in pah1Δ yeast, expansion of the endoplasmic reticulum (with an increase in phospholipid content) facilitates the replication of RNA viruses, which take advantage of the greater membrane surface area and the change in lipid composition. 27 Prolonged exercise (another RM trigger) results in cytokine release, DNA damage, and oxidative stress in skeletal muscle. 28 Although regular exercise is crucial for good health (notably by upregulating the transcription of genes encoding programs involved in fatty acid oxidation and oxidative phosphorylation), overtraining is associated with DNA damage and oxidative stress. 29 Furthermore, lipid droplets accumulate in lipin-1-deficient tissues 5,11,16,30 and might amplify systemic immune responses because the deposits store inflammatory precursors. 31 Accordingly, the inhibition of lipid droplet formation protects mice from inflammatory mediator production and death by endotoxic shock. 32 We showed previously that the lipid droplet count in patients' myoblasts was much lower following treatment with dexamethasone. 11 We believe that our present results open up major perspectives for treating "sporadic" or "epidemic" RM caused by viral infections; these episodes are much more frequent than episodes of hereditary RM. 33 Promising results with dexamethasone have been already described for the treatment of RM variously caused by infection, 33 effort, 34 drugs, and alcohol. 19 Approximately 26 000 cases of RM are reported in the United States annually, and little is known about the recurrence rate. 35 Whatever the cause, all episodes of RM are thought to have the same downstream mechanism: a shortage of energy, dysfunction of the Na/K-ATPase and Ca2+ ATPase pumps, elevated cellular permeability to sodium ions, and a rise in the intracellular calcium concentration. 36 Intracellular calcium activates calcium-dependent proteases and phospholipases, leading to the destruction of myofibrillar, cytoskeletal and membrane proteins, and the release of muscle-cell contents into the circulation. 37 It has been postulated that RM has an inflammatory component specifically due to the release of inflammatory mediators from dying myocytes. 19 Moreover, skeletal muscle is damaged by the myotoxic cytokines released in response to viral infection. 38 Our results have prompted us to treat all cases of severe RM-even in the absence of genetic diagnosis-with corticosteroids, after we have ruled out certain infections in which the drugs are contraindicated. Patient P15 in our study could be an example of this: he might not have died if the corticoids had been immediately administered, but we will never be able to prove it.
Along with the role of inflammation in lipin-1 deficiency, RM involves an energy defect; this is mainly due to the impairment in mitophagy but might also involve abnormal mitochondrial dynamics. In fact, the phosphatidate pool metabolized by lipin-1 interacts with the conserved dynamin-related protein 1. 39 In the future, long-term treatment with TLR9 antagonists, autophagy/ mitophagy clearance inducers, or antioxidants might prevent acute RM. Fasting should be discouraged because it induces autophagy, and early death is observed in starved lipin-deficient flies. 40 In conclusion, our study suggests a simple, inexpensive, easy-to-access treatment for patients with lifethreatening lipin-1 deficiency with RM or with a high risk of RM: oral steroids prior to hospital admission, intravenous corticosteroids in hospital, supportive treatment (hyperhydration), and close monitoring, consistently with the "inflammatory hypothesis" for lipin-1 deficiency. The findings need to be confirmed by further multicenter randomized studies.
AUTHOR CONTRIBUTIONS Caroline Tuchmann-Durand: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. Célina Roda: drafting/revision of the manuscript for content; analysis or interpretation of data. Perrine Renard, Marie Hug de Larauz, Eloise Thevenet, Laure Caccavelli: major role in the acquisition of data. Guillaume Mortamet: major role in the acquisition of data; provision of patient care. Claire-Marine Bérat, Lucile Altenburger, Florence Moulin, Juliette Bouchereau, Anais Brassier, Jean-Baptiste Arnoux, Manuel Schiff, Nathalie Bednarek, Delphine Lamireau, Alexa Garros, Karine Mention, Aline Cano, Jean-Herlé Raphalen, Sylvain Renolleau: provision of patient care. Charles-Henri Cottart, Lionel Finger, Cécile Sergent Brochet: analysis or interpretation of data. Michele Pelosi: revision of the paper. Mehdi Oualha: drafting/revision of the manuscript for content; provision of patient care. Pascale de Lonlay: drafting/ revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data; additional contributions: provision of patient care.