Calcineurin Inhibitor Use and Myoclonus Association. Is There a Clinical Implication? ()
1. Introduction
Movement disorders like tremors and less commonly myoclonus have been reported as adverse reactions of several classes of drugs. Myoclonus is clinically described as a sudden shock-like contraction of a muscle or group of muscles. Myoclonus can further be divided into physiologic and pathologic. Physiologic myoclonus is benign and often seen while falling asleep (hypnic myoclonus) or causing hiccupps (diaphragmatic myoclonus). Pathologic myoclonus can be categorized by three main causes: 1) hypoxia 2) drug toxicity and 3) metabolic disturbances [1].
Drug associated myoclonus usually resolves when the causative drug is removed [2]. For example, anesthetic agents have been described as a cause of myoclonus. In a 2017 study by Jansen et al., drugs associated with inducing/maintaining general anesthesia were the seventh most common drug class to cause myoclonus. Anesthetic agents, however, have a short term exposure to the patient as opposed to the anti-organ rejection drug class, which are a part of the long term management of transplanted patients. The CNI class includes cyclosporin and tacrolimus. Historically, they have been linked to neurotoxicity usually manifested as encephalopathy but myoclonus has also been reported [3] - [8]. Reactions like myoclonus are typically categorized as mild and acceptable given the life-extending outcome the drug is expected to provide [9]. Nevertheless, the association between antirejection drugs and perioperative myoclonus may have long term clinical implications. We sought to quantify the prevalence of myoclonus in heart transplant patients and its clinical relevance using a national dataset.
2. Methods
2.1. Data Source and Cleansing
The United States Nationwide Inpatient Sample (NIS) is the largest all-payer inpatient care database designed to represent a 20% stratified sample of US hospitals or approximately 5 - 8 million hospital stays using data from approximately 1000 hospitals. The NIS has been used in various studies to identify, track, and analyze national trends in healthcare utilization, access, charges, quality, and outcomes [10]. Given the size of the database, it was the preferred source for this study. More details on the design of the NIS are available at http://www.hcup-us.ahrq.gov.
The NIS database from 1 January 2011 to 31 December 2014 was used to analyze data from 842,762 patients. To avoid double representation, patients whose disposition or admission type indicated a transfer to or from another short-term hospital were excluded. Patients with a hospital charge of less than $100 were likely coded incorrectly and were also excluded from the analysis. Similarly, patients with a negative LOS or LOS exceeding 365 days were eliminated from the dataset [11].
The International Classification of Disease, 9th Revision, Clinical Modification (ICD-9-CM) procedure code V42.1 was used to identify the 56,423 patients admitted for heart transplant. This approach has been used previously to estimate in-hospital mortality related to heart transplant in a large sample in the United States [12]. 786,339 patients admitted for CABG surgery were selected as the control group, using the ICD-9-CM procedure codes: 36.10, 36.11, 36.12, 36.13, 36.14, 36.15 and 36.16 [13]. Patients who developed myoclonus during the same hospitalization were determined using ICD-9-CM diagnosis code: 333.2. The presence of underlying epilepsy or hypo-ischemic encephalopathy was determined using ICD-9-CM diagnosis codes 345.0x-345.5x, 345.7x-345.9x, 780.39 and 768.70, respectively and were removed from the study cohort.
2.2. Variables
The variables obtained for this study were: age, sex, and medical comorbidities i.e., hypertension (HTN), diabetes mellitus, coagulopathy, chronic lung disease, and renal failure [10] [14]. This information came from the Agency for Health Research and Quality comorbidity data files. Medical co-morbidities, i.e. dyslipidemia, and atrial fibrillation (Afib) were determined by using the ICD-9CM secondary diagnosis codes (272.40), and (427.31) respectively. ICD-9 secondary codes were used to identify in-hospital complications such as acute kidney injury (584.9, 39.95, 54.98), pneumonia (486, 481, 482.8, and 482.3), urinary tract infection (590.0, 590.9), and sepsis (995.91, 995.92, 996.64, 038, and 999.3). The length of stay, discharge disposition, and in-hospital mortality were determined using the variables ‘‘LOS”, “DISPUNIFORM”, and ‘‘died’’ from Agency for Health Research and Quality comorbidity data files. Prolonged length of stay (PLOS) was defined as length of stay more than 7 days.
2.3. Outcomes
The outcomes of interest were prolonged length of stay greater than seven days, discharge disposition of alive patients to a nursing home and in-hospital mortality.
2.4. Statistical Analysis
The SAS 9.4 software (SAS Institute, Cary, NC) was used to convert NIS database data into weighted counts to generate national estimates, following Healthcare Cost and Utilization Project recommendations. Univariate analysis, Chi-square for categorical, and Analysis of variance (ANOVA) test for continuous variables were performed to identify differences in study variables and outcome endpoints. A statistically significant p-value was considered if <0.05.
This study was exempted from approval by the Texas Tech University Health Sciences Center El Paso Institutional Review Board as NIS is a public database with no personal identifying information.
3. Results
A total of 56,423 patients underwent HT during the study period of which 147 (0.26%) patients had myoclonus while 786,339 patients underwent CABG of which 338 (0.04%) had myoclonus {p < 0.0001}. Two medical comorbidities were different between the two groups in the univariate analysis. A total of 9 patients (6.5%) in the HT + myoclonus group had atrial fibrillation and 131 patients (38.8%) in the CABG + myoclonus group {p < 0.0001}. 94 (64%) patients in the HT + myoclonus group had renal failure and 96 (28.4%) patients in the CABG + myoclonus group {p = 0.0002}. When comorbidities were adjusted for in the data set, the higher rates of myoclonus in the HT group remained statistically significant with a 5-fold difference observed, Table 1 & Table 2.
Patients with myoclonus in both groups had higher complications rates of acute kidney injury compared to patients without myoclonus in their respective groups, Table 3 & Table 4. Similarly, patients with myoclonus in both groups had higher prolonged length of stay compared to patients without myoclonus. Patients in the CABG + myoclonus group had higher discharge disposition of alive patients to nursing home and higher In-Hospital mortality rates than patients without myoclonus, {p = 0.002 and p = 002}. This was not true for patients in the HT + myoclonus group, however there was a trend towards a statistically significant difference for higher in-Hospital mortality {p = 0.08}, Table 5. All data regarding prolonged length of stay, discharge disposition of alive patient to a nursing home and in-hospital mortality were collected from the NIS database.
Table 1. Incidence of Myoclonus in HT and CABG.
Table 2. Comparing basic CABG demographics.
Table 3. Comparing basic HT demographics.
Table 4. Comparing HT vs CABG demographics.
Table 5. Comparing mortality with and without myoclonus.
4. Discussion
In this national representative retrospective study, we found higher rates of myoclonus in patients undergoing HT compared to those undergoing CABG surgery. The difference remained statistically significant after adjusting for potential confounders. Our findings suggest that CNIs may be a contributing factor to myoclonus.
Patients with HT were the preferred organ transplant group selected to study the effect of CNIs because of the high volumes and a similar representative control group (patients undergoing CABG) with comparable surgical and anesthetic conditions. Both cases are done with cardiopulmonary bypass and the anesthetic management is very similar with the exception that CNIs are not administered during CABG. Kidney and liver transplant occur at a higher frequency than HT however finding a comparable control group for those organ transplants would be challenging due to the small number of cases that are done for reasons other than malignancy, i.e. For kidney and liver transplant, a comparable surgery would be hepatectomy or nephrectomy, however, few hepatectomies and nephrectomies are done for non-malignant causes which would therefore limit finding a sizable control group. Given the similar management during both cardiac surgeries, our findings suggest an association between myoclonus and CNIs.
Patients in the HT + myoclonus group had higher complication rates of acute kidney injury compared to those without myoclonus. The nephrotoxic effects of CNIs are well known [15]. Interestingly, patients in the CABG + myoclonus group also had higher complication rates of acute kidney injury despite not receiving CNIs. The authors postulate the hypothesis that there may be an association between renal injury and blood brain barrier dysfunction when other mechanisms lead to acute kidney injury in patients with CABG.
Neurotoxicity has previously been associated to CNIs [4] [6] [8] [9] [16] [17]. In a study published by Adams et al., 13 of 52 orthotopic liver transplants had “fits” (10 grand mal seizure, 2 myoclonus, and 1 focal seizure) [18]. In another study, up to 40% of patients had upper extremity tremors while receiving cyclosporin or tacrolimus after lung or liver transplant [11] [19]. The toxic effects of CNIs are in part due to inhibition of calcineurin by binding to proteins in the cells called immunophilins. Calcineurins are a calcium/calmodulin-dependent protein phosphatase involved in intracellular signaling involved in many physiological processes. High amounts of calcineurins are found in both lymphocytes and within neuronal cells. It is particularly concentrated within the cell body, post synaptic terminals, and axons of neurons. Immunophilins like FKBP12 and cyclophilin are proteins responsible for regulating cellular function including protein folding within the central and peripheral nervous system. It is postulated that the relative amounts and distribution of immunophilins within nerve tissue of each individual may account for the variation and severity of the symptoms to CNIs [20]. Calcineurins themselves exist in several isoforms which all have different degrees of susceptibility to inhibition. This may also account for the differences in rates of neurotoxicity that are observed in these patients [21].
Calcitonin inhibitors are lipophilic agents but contain many aliphatic groups. Under normal physiologic conditions the blood brain barrier limits the access of CNIs to neurons [22]. However, several disease processes (including hypertension, aging, autoimmune disease, and chronic systemic inflammation) have been linked to compromised functionality of the brain blood brain barrier and increased permeability [23] [24] [25] [26] [27]. Interestingly, bovine blood brain barrier models have demonstrated that proinflammatory cytokines and nitric oxide produced by glial cells may contribute to this disruption [28] [29]. Low cholesterol levels have also been shown to elevate Low Density Lipoprotein (LDL) expression on cells. Arachnoid and astrocyte cells primarily express LDL receptors and are vital for maintenance of the blood brain barrier. There is binding of CNIs to LDL receptors to gain entry into these cells exposing them to their toxicity [30]. Patients who have underlying inflammatory states, hypertension, or lower cholesterol levels could therefore be at particular risk for increased blood brain barrier permeability and consequently to CNIs and their toxicity.
In this particular study, although patients in the CABG + myoclonus group had a higher mortality than those without myoclonus this did not hold true for the HT + myoclonus group, Table 5. However, there was a trend toward a statistically significant difference and the authors attributed the low number of HT cases in this national database that underpowered the study to achieve a statistically significant difference. Had the period of study been longer the difference would have been achieved.
Given our findings, the authors suspect that myoclonus may be a clinical indicator of the overall health of these patients, the endothelium health and the health of the blood brain barrier. If chronic disease states can lead to a disrupted and more permeable blood brain barrier, it could likely expose the neural tissue to different caustic agents. Likewise, the compromised function of other organs and systems may also make patients susceptible to neurotoxicity. If the hypothesis of myoclonus as an indicator of poor blood brain barrier function holds true, it stands to reason that even in the absence of CNIs, those patients with chronic conditions or factors that increase the blood brain barrier permeability would still be susceptible to myoclonus and would still have higher rates of mortality through other neurotoxic agents and/or physiologic disturbances. Myoclonus was present in some of these patients in the CABG group despite not being on CNIs.
There are limitations to this study that must be acknowledged. The study was retrospective which presents challenges to mitigating bias and confounding variables. The model used for comparison was surgical patients under general anesthesia. The environment itself introduces variables that cannot readily be controlled or excluded. For example, anesthetic agents like etomidate can also cause myoclonus, however its use is limited to the general anesthesia period. Excluding the use of CNIs for HT, the impact, the anesthetics by themselves would have on the incidence of myoclonus in one group versus the other should be negligible. Patients under general anesthesia have blunted autoregulation of the central nervous system which leads to a more linear relationship between Mean Arterial Pressure (MAP) and Cerebral Blood Flow (CBF). There is the possibility while under general anesthesia increases in MAP including during cardiopulmonary bypass would allow a greater delivery of agents like CNIs across the blood brain barrier due to the increased cerebral blood flow. This enhanced delivery could exaggerate the differences in myoclonus rates between the HT group vs CABG group due to the presence of CNIs. As mentioned previously, nitric oxide appears to play a role in the permeability of the blood brain barrier. There are instances where patients with a history of right heart dysfunction or who are undergoing heart transplant would be placed on inhaled nitric oxide perioperatively. This theoretically could increase the permeability of the blood brain barrier regardless of the general state of health of its endothelium. The authors were not able to exclude the impact inhaled nitric oxide might have had within both groups. However, the use of nitric oxide is not commonly used and should not affect the overall results.
Notwithstanding the limitations of the study, our data indicate significant higher rates of myoclonus within the HT group. Our findings remained after controlling for confounders suggesting that the CNIs and their neurotoxicity may be contributing factors to the higher rates of myoclonus observed in the HT group. A higher mortality rate was observed in patients in the CABG + myoclonus group and a trend in the HT + myoclonus compared to those without myoclonus. The higher mortality observed may implicate the presence of multiple chronic co-morbidities, including dysfunction of the blood brain barrier. More prospective studies are needed to reproduce our findings and to study the significance of myoclonus as a surrogate marker of blood brain barrier function.
Glossary of Terms
Afib—Atrial Fibrillation
ANOVA—Analysis of Variance
CABG—Coronary Bypass Graft
CBF—Cerebral Blood Flow
CNI—Calcineurin Inhibitors
DM—Diabetes Mellitus
HT—Heart Transplant
HTN—Hypertension
ICD-9-CM—International Classification of Disease, 9th Revision, Clinical Modification
LDL—Low Density Lipoprotein
LOS—Length of Stay
MAP—Mean Arterial Pressure
NIS—Nationwide Inpatient Sample
PLOS—Prolonged Length of Stay
Schematic Diagram of NIS Data Sampling Selection