Prevalence of malignant hyperthermia diagnosis in obstetric patients in the United States, 2003 to 2014

Malignant hyperthermia (MH) is a pharmacogenetic disorder of the skeletal muscle triggered by halogenated inhalational agents or the depolarizing muscle relaxant succinylcholine [1, 2]. Exposure to triggering agents can lead to a rapid and uncontrolled calcium release in skeletal muscle cells cytoplasm and to a potentially lethal hyper-metabolic crisis among susceptible patients [3]. Exposure to succinylcholine only, without volatile anesthetics, can also triggers a MH crisis [3, 4]. MH susceptibility is related to defects in the ryanodine receptor 1 (RyR1) or in the Ca_v1.1 channel, two calcium channels located in the sarcoplasmic reticulum membrane, or in the Stac3 protein required for effective colocation of RyR1 and Ca_v1.1 [5]. Defects are caused by mutations in the genes encoding for RyR1, Ca_v1.1 (CACNA1S), or Stac3. Despite the increased awareness of this condition, improved intraoperative monitoring of end-tidal carbon dioxide allowing early detection of MH crisis, and availability of an effective treatment (dantrolene), prognosis of MH crisis is still poor. The 2007–2012 report from the North American Malignant Hyperthermia Registry of the Malignant Hyperthermia Association of the United States indicates a case fatality rate of nearly 10%, with increases in MH-associated morbidity and mortality when dantrolene administration is delayed [3, 6, 7].

MH susceptibility is a rare disorder. Reported prevalence of MH diagnosis in studies using hospital discharge records from administrative data ranges from 0.18 per 100,000 (95% confidence interval (CI), 0.12–0.25) in ambulatory surgery center patients to 0.96 (95% CI, 0.75–1.41) in surgical inpatients [8, 9]. The low frequency of MH crises and the cost of stocking dantrolene have given rise to concern about the cost-benefit of the recommendation of the Malignant Hyperthermia Association of the United States (MHAUS) that dantrolene be made immediately available (for administration within 10 min) in operating room areas, especially in facilities with a low utilization of general anesthesia and triggering agents such as maternity units [10, 11]. MHAUS is a patient safety and advocacy organization and its recommendations are used by accrediting agencies, such as the Joint Commission, to assess preparedness for an MH event during survey visits. However, no study has specifically examined the prevalence of MH susceptibility in obstetric patients [12].

Using a nationally representative sample of community hospital discharges records in the United States between 2003 and 2014, we estimated the prevalence of MH diagnosis in delivery-related discharges and assessed patient- and hospital-level factors associated with the prevalence of MH diagnosis.

During the 12-year study period, 9,892,712 delivery-related discharges were identified in the NIS (weighted number 47,178,322) (Fig. 1). Of them, 3,181,978 (32.2%) were cesarean delivery discharges (weighted number 15,175,127).

A diagnosis of MH was recorded in 46 delivery-related discharges (weighted number 215), yielding a prevalence of MH in obstetrics of 0.46 per 100,000 delivery-related discharges (95% confidence interval (CI), 0.40–0.52) (Fig. 2). In cesarean delivery discharges, a diagnosis of MH was recorded in 26 cases (weighted number 123), yielding a prevalence of 0.81 per 100,000 (95% CI, 0.67–0.97). In vaginal delivery discharges, a diagnosis of MH was recorded in 20 cases (weighted number 92), yielding a prevalence of 0.29 per 100,000 (95% CI, 0.23–0.35). The prevalence of MH in cesarean delivery discharges was significantly higher than in vaginal delivery discharges [P-value < 0.001; crude OR 2.83; 95% CI, 2.16 to 3.71; aOR 2.88; 95% CI, 2.19 to 3.80].

Prevalence of MH in delivery-related discharges according to patient characteristics, hospital characteristics, and time period is presented in Table 1. On the bivariate level, the prevalence of MH diagnosis was significantly lower in Hispanics than in non-Hispanic Whites and significantly higher in the Midwest and South census regions than in the Northeast census region, and in rural hospitals than in urban hospitals. No significant difference was observed across age groups, comorbidity indexes groups, or time periods. Multivariable adjustment revealed two factors associated with MH diagnosis: race/ethnicity (Hispanic versus non-Hispanic White, aOR 0.47; 95% CI, 0.29 to 0.76) and geographic census region (South versus Northeast, aOR 2.44; 95% CI, 1.50 to 3.96).

This study was designed to help close a knowledge gap in the epidemiology of MH. Specifically, we provide an estimate of the prevalence of MH diagnosis in the obstetric patient population using a nationally representative sample over a 12-year period. The prevalence of 0.81 per 100,000 (95% CI, 0.67 to 0.97) in cesarean delivery discharges is similar to the one reported in non-obstetrical surgery inpatients. Using New York State administrative data from 2001 to 2005, Brady et al. reported a prevalence of MH diagnosis in inpatient surgical discharges of 0.96 (95% CI, 0.67 to 1.24) [8]. The prevalence reported in the present study is lower than the rate previously reported using NIS data from 2000 to 2005 for the general inpatient population (1.12 per 100,000, 95% CI, 1.08 to 1.17) [18]. The discrepancy is likely due to the difference by the study populations as our analysis is restricted to obstetric patients and the prevalence of MH diagnosis in women is only about a third of the one in men [8, 19].

We identified several patient- and hospital-level characteristics associated with a higher prevalence of MH diagnosis. We observed a lower prevalence of MH diagnosis in Hispanics compared with non-Hispanic White patients. To our knowledge, this finding has not been reported before and if confirmed, warrants further investigation to elicit the underlying mechanisms. The lower prevalence of MH diagnosis in Hispanic obstetric patients is unlikely due to a lower utilization of general anesthesia and exposure to triggering agents because two previous studies have reported a higher utilization of general anesthesia in minority patients, including Hispanics [20, 21]. We also observed a geographic variation in the prevalence of MH diagnosis, with an elevated prevalence in the South region. Geographical variation is explained by the concentration of MH families in a given geographic area [22]. Rosero et al. also observed a higher prevalence of MH diagnosis in the South census region. Last, Rosero et al. reported an increase in the prevalence in MH diagnosis between 2001 and 2005 [18]. We do not confirm this trend over the longer 2003–2014 time period. The study by Rosero started just 2 years after the introduction of the ICD-9-CM code for MH in 1998 and the increase in this study may merely be related to the novelty of the code.

The prognosis of MH crisis remains poor with a fatality rate of nearly 10% and MH-associated morbidity and mortality increased when dantrolene administration is delayed [6, 7]. Similar poor prognosis is observed in studies using administrative data. In these studies, discharges recording a diagnosis of MH have a fatality rate ranging from 12 to 22% [8, 18]. In our study, we identified no death in discharges with a diagnosis of MH. A possible explanation for this finding is the young age of the patients analyzed and the lack of associated comorbidity; 98% of MH discharges were under 40 and only 0.3% had a Charlson comorbidity index greater than 1. By contrast, Brady et al. reported a 22% fatality rate in discharges with a diagnosis of MH but 49% of the MH discharges were older than 45 and 46% had a Charlson comorbidity index greater than 1 [8].

The Malignant Hyperthermia Association of the United States recommends that “dantrolene must be available for all anesthetizing locations where MH trigger agents are used.” This recommendation also applies to facilities where volatile agents are not available or administered, and succinylcholine is only stocked on site for emergency purposes [3]. The cost of an MH cart in maternity unit for responding to MH crisis has been estimated at about $2000 per year, including cost for the dantrolene itself, cost for the cart and non-dantrolene items, and cost for MH cart maintenance [10]. Ho et al. have recently questioned this recommendation based on cost-benefit analysis of stocking dantrolene in maternity units [10]. The low incidence of MH crisis and the low utilization of general anesthesia and triggering agents in obstetrics would not make stocking dantrolene policy cost-beneficial. Similar concerns have also been raised in ambulatory surgery centers but two recent studies support the dantrolene stocking policy in these centers [3, 23]. Of note, the prevalence of MH susceptibility of 0.81 per 100,000 cesarean deliveries (or 1 in 123,456 cases) reported in our study, similar to the prevalence reported in non-obstetrical surgery inpatients, suggests that stocking dantrolene in maternity units is justified. In our opinion, the low annual cost of a MH cart to treat a MH crisis in a maternity unit ($2000 per year) and the prevention of maternal death from a potentially treatable complication provide adequate justifications for stocking dantrolene in maternity units. This is particularly relevant in the current US context of implementing nationwide initiatives to reduce preventable severe maternal morbidity and mortality [24].

It is worth noting the limitations with our study. First, the NIS is a stratified sample of hospital discharges records and our study sample includes only 20% of childbirths during the study period. Second, identification of MH diagnosis relies on the accuracy of coding in discharges data and is susceptible to misclassification because of error or absence of coding. In a previous study evaluating the accuracy of MH diagnosis in hospital discharge records, we found that the ICD-9-CM code 995.86 (“Malignant hyperthermia due to anesthesia”) corresponded to a MH crisis in 24% of the cases, to a MH history in 47%, and to fever unrelated to MH in 24% [25]. Therefore, most of the cases with an MH diagnosis identified in the present study should be regarded as MH susceptibility rather than incident MH episodes. Lack of information on anesthetics exposed or dantrolene administration in the NIS dataset precludes us from identifying incident MH episodes. In other words, an ICD-9-CM-based approach favors sensitivity over specificity. Clinical registries such as the North American Malignant Hyperthermia Registry represent more specific approaches but also suffer from limitations such as the voluntary declarations of MH crisis with a risk of under declaration and underestimation of the prevalence of MH crisis. However, these two approaches should not be viewed as antagonistic but as complementary. Third, our unit of analysis was the discharge record and not the patient. So, a patient with multiple hospitalizations during this 12-year study could be included more than once in our analysis. However, the National Inpatient Sample is a 20% sample of discharges records, which decreases the likelihood of a patient being sampled and included multiple times. To further assess the potential impact of repeated admissions of MH-susceptible patients, we checked individually linked hospital discharge records for 2,851,697 obstetric patients in New York State from 2003 to 2014; of the discharges with a diagnosis of MH, none involved repeated admissions. Fourth, this study was limited to obstetric patients because of the recent controversy surrounding dantrolene stocking policy in maternity units. There are other settings where triggering agents may be used, especially succinylcholine, such as emergency departments or intensive care units and where stocking dantrolene could be suggested. To date, no study has specifically examined the prevalence of MH diagnosis in these locations. Fifth, a diagnosis of MH in women who underwent vaginal delivery may appear surprising because a general anesthetic with exposure to MH triggering agents is not required for an uncomplicated vaginal delivery. However, a general anesthetic may be required after a vaginal delivery in some circumstances such as the removal of a retained placenta or an exploration of the uterine cavity because of a postpartum hemorrhage. Moreover, as indicated in our previous study, most of the MH diagnoses in hospital discharge records refer to positive MH history and susceptibility rather than acute MH episodes [25]. These two factors may explain why the prevalence of MH diagnosis in vaginal delivery was much lower than in cesarean delivery (0.29 per 100,000 versus 0.81 per 100,000. Finally, the NIS does not capture all childbirths occurring in the country. For example, in 2014, the NIS included 3,828,643 childbirths which corresponds to 96% of the 3,988,076 births reported by the US National Center for Health Statistics. The “missing” 4% can be partially explained by out-of-hospital births (births at home or in a birth center), which accounted for 1.5% of live births in 2014.

The prevalence of MH-susceptibility is about 1 in 125,000 in cesarean deliveries, similar to the prevalence reported in non-obstetrical surgery inpatients. Results of this study suggest that stocking dantrolene in maternity units is justified.