Anesthetic hypersensitivity in a case-controlled series of patients with mitochondrial disease

Abstract

Background.

Children with mitochondrial disease undergo anesthesia for a wide array of surgical procedures. However, multiple medications used for their peri-operative care can affect mitochondrial function. Defects in function of the mitochondrial electron transport chain (ETC) can lead to a profound hypersensitivity to sevoflurane in children. We studied the sensitivities to sevoflurane, during mask induction and maintenance of general anesthesia, in children presenting for muscle biopsies for diagnosis of mitochondrial disease.

Methods.

In this multi-center study, ninety-one children, ages 6 months-16 years, presented to the OR for diagnostic muscle biopsy for presumptive mitochondrial disease. General anesthesia was induced by a slow increase of inhaled sevoflurane concentration. The primary endpoint, end tidal sevoflurane necessary to achieve a Bispectral index (BIS) of 60, was recorded. Secondary endpoints were maximal sevoflurane used to maintain a BIS between 40–60 during the case, and maximum and minimum heart rate and blood pressures. After induction, general anesthesia was maintained according to the preferences of the providers directing the cases. Primary data was analyzed comparing data from patients with complex I deficiencies to other groups using nonparametric statistics in SPSS v.27.

Results.

The median sevoflurane concentration to reach of BIS of 60 during inductions [End Tidal (ET) Sevoflurane % (BIS=60)] was significantly lower for patients with complex I defects (0.98%; 95%CI, 0.5–1.4) compared to complex II ((1.95% ; 95%CI, 1.2–2.7, p <.001)), complex III (2.0% ; 95%CI, 0.7–3.5 p <.001), complex IV ((2.0% ; 95%CI, 1.7–3.2; p <.001), and normal groups (2.2% ; 95%CI, 1.8–3.0, p <.001). The sevoflurane sensitivities of complex I patients did not reach significance when compared to patients diagnosed with mitochondrial disease but without an identifiable ETC abnormality (p=0.172). Correlation of complex I activity with ET Sevoflurane % (BIS=60) gave a Spearman’s coefficient of 0.505 (p<0.001). The differences in sensitivities between groups were less during the maintenance of the anesthetic than during induction.

Conclusions.

The data indicate that patients with complex I dysfunction are hypersensitive to sevoflurane compared to normal patients. Hypersensitivity was less common in patients presenting with other mitochondrial defects or without a mitochondrial diagnosis.

Background

Many anesthetic agents have been shown to have inhibitory effects on the function of isolated mitochondria from both mammals and invertebrates¹. Studies have primarily indicated that anesthetics inhibit the function of the mitochondrial electron transport chain (ETC)^2–4. The ETC consists of four protein complexes (complexes I-IV) which move electrons from electron donors (primarily NADH or succinate) to oxygen to generate water. The flow of electrons generates a transmembrane potential that powers a fifth complex (complex V) to form ATP (adenine triphosphate) from ADP (adenine diphosphate). The volatile anesthetics are specific and potent inhibitors of complex I of the ETC, which may actually contribute to their mode of anesthetic action^3–7.

Mitochondrial disease is recognized as an important cause of a wide range of neurological, cardiac, and endocrine diseases^8,9. The incidence of disorders of the respiratory chain alone is estimated to be about 1 to 1.5 per 5,000¹⁰. It has become increasingly common for children with mitochondrial defects to undergo surgery and anesthesia¹ making it likely that pediatric anesthesiologists will care for such patients. However, diagnosis and optimized care for children with mitochondrial diseases remains problematic. Surgical procedures usually involve a general anesthetic, but there is little consensus as to the ideal medications to use^1,11. Many different anesthetic techniques have been successfully employed for patients with mitochondrial disease^1,12,13. However, there are also reports of serious complications, including fatal outcomes, during and following anesthetic exposure^14–18. Since mitochondrial disease can result from widely varied defects in the mitochondrion, it is unlikely that a single anesthetic approach will be ideal for all mitochondrial patients. No prospective studies have established the ideal anesthetic for mitochondrial patients.

The potential for additional mitochondrial inhibition in an already compromised patient with known mitochondrial dysfunction is balanced by the safety benefit of their favorable pharmacokinetics, namely their quick elimination when the anesthetic is terminated. In contrast, propofol is known to inhibit several mitochondrial functions, as noted in multiple studies of propofol infusion syndrome^1,19–21. As a result, the use of propofol infusions is generally discouraged in patients with mitochondrial defects¹.

Data in both human and animal models indicate that some mitochondrial patients may be hypersensitive to the volatile anesthetics^11,22–24. Genetic mutants in Caenorhabditis elegans^22,25, Drosophila²⁴, and in mice²³, have all shown hypersensitivity to volatile anesthetics. Furthermore, we previously demonstrated in a small cohort that some children with complex I dysfunction are extremely sensitive to sevoflurane¹¹. Here, we extend these data in a prospective multicenter study to determine the characteristics of sensitivity to sevoflurane in patients with mitochondrial defects.

We hypothesized that patients with a complex I defect would be more sensitive to sevoflurane during anesthetic induction than would patients with normal complex I function. Our primary objective was to determine if patients with mitochondrial complex I defects required lower End Tidal (ET) sevoflurane concentrations to reach a Bispectral Index (BIS) of 60 during anesthetic induction than did patients with different mitochondrial deficiencies or with no mitochondrial abnormalities. We also characterized the clinical course of short surgical cases done with sevoflurane as the primary anesthetic.

Methods.

The study was approved by the IRB committees at University Hospitals of Cleveland, the Cleveland Clinic, and Seattle Children’s Hospital. 2) written informed consent was obtained from all subjects, a legal surrogate, the parents or legal guardians for minor subjects. Most patients were induced by one of three anesthesiologists (VCH, JN, PGM) experienced in the care of mitochondrial patients, using an established protocol for induction (below) followed by standard medical care. The specific biochemical defects in those children who were positive for an electron transport chain (ETC) defect was not known prior to biopsy. Therefore, providers were blinded as to the etiology of disease at the time of biopsy.

Patients.

Children aged six months to 16 years with a presumptive diagnosis of mitochondrial disease who were scheduled for a diagnostic muscle biopsy were eligible for study (Figure 1). Laboratory testing was done only as indicated for their clinical care. All muscle biopsies were analyzed by a single laboratory (CLH). Patients were continued on their usual medications preoperatively, and no preoperative sedatives were given. A BIS© (Bispectral Index, Aspect, Norwood, MA, USA) monitor was placed on each child prior to induction, as well as standard intraoperative monitors. Patients were excluded from the study for preoperative use of sedatives, preoperative BIS<80, preoperative complications that would either potentially affect mental status or make a slow anesthetic induction problematic.

Figure 1.

Study participant flow diagram from enrollment through muscle biopsy results. Of the 91 patients that were initially eligible for the study, 11 were excluded for reasons that included unexpectedly low pre-anesthesia induction BIS values less than 80 (n=4) and requirement for preoperative anxiolysis with midazolam (n=2). Of the 80 patients that underwent muscle biopsies and were included in the study, 48 had test results consistent with mitochondrial disease. The resulting distribution of mitochondrial complex deficiencies is shown.

Anesthetic Management

All anesthesiologists used the following standardized procedure for induction. Patients were scheduled as the first case of the day and were NPO for fluids for 2–3 hours. Anesthesia was induced using a tight-fitting mask with continuous monitoring of end tidal gas concentrations. During induction, concentrations of inspired sevoflurane were begun at 0.2–0.4%, and slowly increased by 0.2–0.4% every one minute or as tolerated by the patient, until a BIS of 60 or less was reached. Inspired sevoflurane was increased only after the end tidal concentration of sevoflurane was constant for at least one minute. Each induction (except for the patients requiring very low doses of sevoflurane) took 6–10 minutes. This slow induction was used to approximate steady state levels of sevoflurane balanced against an acceptable induction time for the patients. Since we increased sevoflurane concentrations at a predetermined rate (sevoflurane was begun at 0.2–0.4%, and slowly increased by 0.3–0.5% every one minute), patients with higher requirements for sevoflurane to reach a BIS of 60 had a correspondingly longer induction times.

Once a patient reached a BIS of 60, loss of consciousness was confirmed by lack of response to voice and by loss of eyelash reflexes. The end tidal concentration of sevoflurane required to reach a BIS of 60 was recorded. BIS was held at 40 to 60 for 5 minutes while the IV was started and the patient was prepared for surgery. The maximal end tidal concentration of sevoflurane necessary to maintain a BIS reading of 40–60 during the procedure was also recorded for each patient.

Intravenous normal saline with 2.5–5% dextrose was started (except in those patients on a ketogenic diet) and run at the appropriate rate to replace the patient’s fluid deficit over 3 hours. No patients received muscle relaxants. After induction, anesthesia was primarily maintained with sevoflurane but with ancillary pharmacologic support in most cases. Airway management, intraoperative use of opioids or local anesthetics, or use of peripheral nerve blocks were left to the discretion of the anesthesiologist. Room temperature adjustments, heating lamps, and forced air warming blankets were used to maintain the patient’s temperature between 36.5 and 37.5°C.

Mitochondrial Diagnoses.

Assays of mitochondrial respiratory capacity via rates of oxidative phosphorylation and/or electron transport chain (ETC) enzymatic activity were performed by the Clinical Pharmacology Laboratory in the Center for Inherited Disorders of Energy Metabolism (CIDEM; Case Western Reserve University School of Medicine)²⁶. ETC enzyme activities were evaluated in skin fibroblasts, fresh skeletal muscle, or in solubilized mitochondria from fresh skeletal muscle.

Patients with functional defects in their muscle biopsy that could be assigned to specific complexes of the ETC were identified. For this study, diagnoses of a complex deficiency was made when the respective complex function was less than half of a large cohort of historical controls. These were not the same criteria as used for medical diagnosis of a complex defect, which were more stringent. Patients for whom ETC and oxidative phosphorylation assays added no new diagnostic information, but had previous genetic or histologic markers for mitochondrial disease, were classified as no ETC defect. They were however considered as patients with mitochondrial disease. Patients in whom no defects were identified in muscle biopsy, and had no other laboratory or historical data diagnostic of mitochondrial disease, were classified as normal. However, they each had a medical history sufficiently suggestive of mitochondrial disease to lead to a muscle biopsy. We did not include a non-disease group. The characteristics of the induction were not known to the laboratory personnel performing the biochemistry studies.

Statistical Considerations.

The primary outcome for the study was the sevoflurane concentration needed to reach a BIS value of 60 during a controlled induction by face mask. Secondary outcomes were maximal sevoflurane concentration and vital signs during the surgical case. Hypersensitivity was defined as an end tidal concentration of less than 1.5% to achieve a BIS of 60 during induction of general anesthesia. We are unaware of a standard definition for hypersensitivity, so relied on an effect size that we felt could be easily recognized clinically (a change of 1% ET_SEVO from the reported MAC for sevoflurane (2.5%))²⁷. In addition, we relied on the earlier results we obtained from a pilot study with similar patients where an apparent demarcation at ~1.6% sevoflurane existed between complex I patients and others¹¹. We present median values with 25% and 75% percentiles in the Figures 2 and 3. Whisker marks are the 95% percentiles.

Figure 2.

Distributions of sevoflurane concentrations required to reach a BIS=60 during induction in patients with mitochondrial defects. The medians (middle line), 25% and 75% quartiles (boxes), and 95% confidence intervals (error bars) of the sevoflurane concentrations necessary to reach a BIS of 60 during induction. Median ET Sevoflurane % for a BIS of 60 is significantly lower for complex I patients than any other group (p<.001 in all cases) except for mitochondrial patients with no identified ETC abnormality. Median ET Sevoflurane % (BIS=60) for each group, p values for groups compared to complex I patients and r value for effect sizes are as follows: Patients with complex I defects (0.93%; 95%CI, 0.37–1.5), complex II(1.95% ; 95%CI, .74–3.15, p <.001; r=0.63), complex III (2.48% ; 95%CI, 1.95–4.75; p <.001; r=0.67), complex IV(2.4%; 95%CI, 1.3–3.5; p <.001, r=0.75), mitochondrial patients without a clear ETC abnormality (1.7%; 95% CI 0.98–2.2; p=0.172), and normal groups (2.0%; 95%CI, 1.3–2.5; p <.001, r=0.51). ^★ indicates p<0.001 for a comparison to values from patients with a complex I deficiency.

Figure 3.

Distribution of maximal sevoflurane concentrations required for maintenance of surgical anesthesia necessary for each patient in Figure 2. General anesthesia was maintained with sevoflurane plus ancillary medications as described in Supplementary Table S2. The medians (middle line), 25% and 75% quartiles (boxes), and 95% confidence intervals (error bars) of the sevoflurane concentrations necessary to maintain a BIS of 40–60 for the patients. ^★ indicates p<0.01 for a comparison to values from patients with a complex I deficiency. Mean ET Sevoflurane % (BIS=40–60) for each group and precise p values for groups compared to complex I patients are as follows: Patients with complex I defects (2.4%; 95%CI, 1.8–3.0), differ significantly only when compared to the normal group (2.8%; 95%CI, 2.0–3.6; p =.001, r=0.50). When compared to patients with complex II((2.6% ; 95%CI, 1.8–3.3, p =.101), complex III ((2.6% ; 95%CI, 1.4–4.3) p =.102), complex IV((2.8%; 95%CI, 1.75–3.85; p =.186), and mitochondrial patients without a clear ETC abnormality ( 2.5%; 95% CI 1.0–3.4; p=0.096) statistical significance was not reached.

The data for all patients was evaluated to satisfy a normal distribution. While the data from complexes I, III, mitochondrial diagnosis without ETC abnormality, and normal mitochondria, each appeared to satisfy such a distribution, the data for complexes II and IV did not. In addition, the standard deviations were quite varied. Thus, we used nonparametric analyses, a Kruskal-Wallis test comparing all groups followed by Mann-Whitney tests between groups. We assessed the relationship between mitochondrial complex I activity and induction concentration of sevoflurane (using BIS=60 as an endpoint) while controlling for the confounding effects from age, BMI and gender using multiple linear regression analysis. The relationship was modelled assuming a linear relationship between the mitochondrial complex I activity and induction concentration of sevoflurane as well as after logarithmic transformation of the mitochondrial complex I activity. We also used the nonparametric Spearman test for correlation of complex I activity with ET Sevoflurane (BIS=60). Finally, effect sizes were calculated from the Mann-Whitney test using r=z/(n)^1/2 and reported in the Results as r values. An r=0.5 means that 50% of the difference can be accounted for by the variable being considered. We compared complex I results to each of the other groups, giving five comparisons. We then used a p<0.01 as our cutoff for significance with a Bonferroni correction for the five comparisons. Thus our final significance cutoff was 0.002. SPSS v. 27 was used for all statistical analyses.

A formal sample size calculation was not performed. All patients known to the authors to be presenting for muscle biopsy to rule out mitochondrial disease were approached for the study. Three complications limited the number of families approached. First, one of the authors (PGM) is not a full-time clinician. Secondly, PGM and MMS moved during the study to an institution performing fewer such biopsies than their previous hospital. Finally, the incidence of muscle biopsies performed decreased as exome sequencing became more readily available as a diagnostic aid. We approached 91 patients for enrollment (Figure 1). Eleven (11/91) patients were withdrawn (see Figure 1 and Results for descriptions of causes).

Results.

A total of 91 patients with suspected mitochondrial disease were initially enrolled in the study (Table 1, Figure 1). All patients had muscle biopsies; none of the four patients removed for low BIS values had abnormal mitochondrial function studies. In total, 48 patients retained a diagnosis of mitochondrial disease while 32 patients did not. No patient had serious complications during or after the surgeries; all patients were discharged from the hospital on the day of surgery. Detailed demographics are found in Supplemental Tables S1 and S2.

Table 1.

Summary of demographics of patients with different mitochondrial diagnoses.

Mitochondrial Deficiency	N	Ages (M) (S.D.)	Gender F/M	BMI (S.D.)	ET Sevoflurane Induction	ET Sevoflurane Induction
Complex I	14	66.4 (50)	4/10	18.8 (6.1)	0.975 (0.5–1.4)	2.0 (1.9–2.7)
Complex II	4	71.8 (80)	2/2	19.8 (5.7)	1.95 (1.2–2.7)	2.6 (2.2–3.0)
Complex III	15	79 (62)	7/8	19.3 (5)	2.0 (.71–3.5)	2.4 (1.9–4.9)
Complex IV	5	75 (53)	2/3	18.5 4.9	2.0 (1.7–3.2)	3.2 (2.1–3.5)
No ETC Abnormality	10	101 (67)	5/5	19.4 (7.3)	1.7 (.44–2.8)	2.95 (2.1–2.95)
Non-Mitochondrial	32	75 (55)	15/17	20.3 (12.9)	2.15 (1.3–3.0)	3.2 (2.4–4.4)

Anesthetic Induction.

The distribution of sevoflurane concentrations necessary to reach a BIS of 60 are shown in Figure 2. Fourteen of fourteen (14/14) patients with complex I defects achieved a BIS of 60 with an induction concentration of sevoflurane of less than 1.5%. However, 2/4 patients with complex II defects and 3/15 patients with complex III defects also achieved a BIS of 60 with low concentrations of sevoflurane during induction. In addition, 4/10 with currently undiagnosed mitochondrial defects also were hypersensitive to sevoflurane. None of the 5 patients with complex IV defects were hypersensitive to sevoflurane. Of the children who were not diagnosed as having mitochondrial disease (and thus served as controls), one of thirty-two (1/32) patients demonstrated hypersensitivity to sevoflurane. Comparison of all groups for induction using the Kruskal-Wallis test gave a p<0.001, thus pairwise comparisons were made between groups. The median sevoflurane concentration (Figure 2) to reach of BIS of 60 during inductions was significantly lower for patients with complex I defects (0.95%; 95%CI, 0.37–1.5) compared to complex II ((1.95%; 95%CI, 0.74–3.15), p <.001; (r=0.63; 95%CI, 0.27–0.93)), complex III ((2.4%; 95%CI, 1.95–4.75) p <0.001; (r=0.67; 95%CI, 0.30–0.99)), complex IV ((2.4%; 95%CI, 1.3–3.5); p <0.001, (r=0.75; 95%CI, 0.41–1.02)), and normal groups ((1.9%; 95%CI, 1.3–2.5); p <0.001,( r=0.51; 95%CI, 0.36–0.65)). The results for complex I failed to reach significance when compared to the results for mitochondrial patients without a clear ETC abnormality (1.6%; 95% CI, 0.8–3.0; p=0.172).

Anesthetic Maintenance.

During maintenance of general anesthesia, smaller differences were noted between groups for intraoperative maximal concentrations of sevoflurane (Figure 3) than for induction of general anesthesia. During the course of surgery, other medications, most often fentanyl or regional analgesia, were given to the patients which limited the ability to interpret sensitivities across patient groups. Comparison of all groups for maintenance using the Kruskal-Wallis test gave a p=.049, thus pairwise comparisons were made between groups. The median sevoflurane concentration (Figure 3) to maintain a BIS of 40–60 during the maintenance phase of surgery was significantly lower for patients with complex I defects (2.4%; 95%CI, 1.8–3.0) only when compared to the normal group (2.8%; 95%CI, 2.0–3.6; p =.001, r=0.50). When compared to patients with complex II((2.6% ; 95%CI, 1.8–3.3, p =0.101), complex III ((2.6%; 95%CI, 1.4–4.3) p =0.102), complex IV((2.8%; 95%CI, 1.75–3.85; p=0.186), and mitochondrial patients without a clear ETC abnormality (2.5%; 95% CI 1.0–3.4; p=0.096), the differences failed to reach significance.

Association between ET Sevoflurane (BIS=60) and complex I activity

We next determined whether anesthetic sensitivity for induction was correlated with complex I activity of the patients. For this correlation, we compared the induction concentrations of sevoflurane with the normalized complex I activities for all patients (Figure 4). We used either complex I activity determined either by complex I activity determined by enzymatic assays (ETC) or by oxidative phosphorylation when an ETC value was not available. In each case, the activities were normalized to midpoint of the concurrent normal ranges. Sixteen of twenty-four (16/24) patients that required less than 1.5% sevoflurane for induction to reach a BIS of 60 had less than 75% of normal complex I function. Spearman’s correlation coefficient comparing ET Sevoflurane (BIS=60) with normalized complex I activity was 0.505 with a 2-tailed significance of p<0.001. Regression analysis using a linear model showed a moderate correlation (R² = 0.202) with a standardized coefficient of 0.450 (95% CI 0.404–1.093, p<0.001). After allowing for the confounders of age, BMI and gender, regression analysis comparing ET the standardized coefficient was 0.471 (95% CI 0.414–1.134, p<0.001). When sevoflurane (BIS=60) was compared with the logarithm of complex I activity, the correlation improved (logarithmic transformation of the mitochondrial complex I activity; R² = 0.258) compared to the linear fit although the correlation remained moderate. Regression analysis using the logarithmic model gave a standardized coefficient of 0.393 (95% CI 0.305–1.015, p<0.001). After allowing for the confounders of age, BMI and gender, regression analysis gave a standardized coefficient of 0.471 (95% CI 0.342–1.062, p<0.001). However, since both relationships are only moderate the results should be considered preliminary.

Figure 4.

The effective concentration (ET Sevoflurane % (BIS 60)) during induction of anesthesia in individual patients versus the normalized complex I activity for that patient. Complex I values were determined either by ETC assays or oxidative phosphorylation studies. In each case the values were normalized relative to the midpoint of normal complex I dependent activity for the test used to determine complex I function. Lines are best fits using a linear regression (solid line) or logarithmic regression (dashed line). Spearman’s correlation coefficient was .505 with a 2-tailed significance of p<0.001. Regression analysis comparing ET Sevoflurane % (BIS=60) with normalized complex I activity best fit a logarithmic relationship (R² = .258) compared a linear fit (R² = .201). Text deleted here.

Discussion

We have extended our previous pilot study¹¹ to confirm that some patients with mitochondrial dysfunction have increased sensitivity to sevoflurane, as defined by reaching a BIS^© of 60 at sevoflurane concentrations below 1.5%. Our primary finding was that all children with complex I disease were hypersensitive to induction with this volatile anesthetic (Figure 2). The hypersensitivity supports the use of some form of electroencephalogram (EEG) monitor to measure anesthetic depth in patients with mitochondrial disease to help avoid exposure to unnecessarily high concentrations of sevoflurane. In our experience, the slow inductions used in this are not common in clinical practice. We think that that use of EEG monitoring can lessen the likelihood of reaching a state of burst suppression during an apparently normal induction. However, increased sensitivity during induction did not always indicate a decreased concentration of sevoflurane necessary for immobility during surgery in these patients (Figure 3). Once a BIS of 60 or less was reached, maintenance of anesthesia for surgery was conducted using sevoflurane supplemented with ancillary medications. Five of the fourteen complex I patients required concentrations of sevoflurane >2%, in addition to ancillary medications, to eliminate movement or an increase in BIS, heart rate, or blood pressure during surgery. We hypothesize that this reflects the difference in endpoints, a BIS measure of unconsciousness versus a steady state MAC for painful stimulus. Since maintenance phases of the surgeries were not protocolized, meaningful comparisons between groups were not possible.

Volatile anesthetics have been shown to specifically depress complex I-dependent oxidative phosphorylation^4,6,28; other steps within the electron transport chain were not inhibited by clinical concentrations of volatile anesthetics. In addition, prior work from our laboratory and others^22–25 has shown that sensitivity to volatile anesthetics of nematodes, mice and flies is strongly dependent on complex I function. It was suggested that this hypersensitivity is the result of decreased neurotransmitter recycling in glutamatergic neurons^5,29. Therefore, the hypersensitivity of patients with complex I defects was not unexpected. However, we also observed hypersensitivity in some patients with defects in other complexes. Complexes I, III and IV exist in supercomplexes³⁰ and have been shown, in model systems, to functionally interact, e.g. defects in complexes III and IV can alter the function of complex I^31,32. It is possible that the defects seen in complex III patients may cause hypersensitivity via effects on complex I. Finally, as can be inferred from Figure 4, this measure of anesthetic sensitivity, while correlating moderately with complex I activity, is not uniquely dependent on its activity. Even with normal complex I activities, there is a wide range of sensitivities to sevoflurane for induction, indicating that other targets must also play a role. However, it is remarkable that even in those patients with striking neuromuscular disease from mitochondrial defects in complexes II, III, or IV, there was not a uniform increase in sensitivity to sevoflurane.

Others have studied anesthetic regimes of patients with mitochondrial disease. Driessen et al., published a retrospective case series of 122 patients with muscle biopsy-confirmed mitochondrial defects¹². They found no relationship between the anesthetic used and any perioperative complications but did not study sensitivity during induction. Footitt et al. published a smaller retrospective series of 38 pediatric patients reporting that patients with mitochondrial disease may undergo a brief general anesthetic using volatile anesthetics or propofol without serious anesthesia-related adverse events³³. Smith et al. published a retrospective series of 26 patients with mitochondrial disease having 65 general anesthetics³⁴. They found no correlation between the type of anesthetic and subsequent complications. However, 4/26 patients were reported to have a hemodynamic complication although no discussion of the type of complication nor of the severity was given. None of these three articles reported anesthetic concentrations or doses necessary for induction or maintenance of anesthesia. In each report, the unknown molecular causes for mitochondrial disease limits the ability to generalize the results to all mitochondrial patients. However, this study adds to the data suggesting that volatile anesthetics, delivered with appropriate monitoring, can be used safely in this population at least in short cases.

Hypersensitivity of animals with complex I dysfunction to volatile anesthetics has been clearly demonstrated across the animal kingdom. Investigations in Caenorhabditis elegans, Drosophila melanogaster, and Mus musculus have each demonstrated that complex I defects lead to striking hypersensitivity to different volatile anesthetics^23–25. Mutations affecting other complexes did not alter anesthetic sensitivity in C. elegans^31,32,35. The degree of inhibition of complex I-dependent oxidative phosphorylation strongly correlated with the effect on anesthetic sensitivity in a wide survey of C. elegans mutants. In human patients we do not know if extent of complex I dysfunction can generally predict degree of anesthetic hypersensitivity. However, the data shown in Figure 4 suggest that degree of complex I activity per se may relate to sensitivity to sevoflurane during induction when no ancillary medications are given.

The intent of this study was not to determine a precise minimum alveolar concentration (MAC) to produce unconsciousness in mitochondrial patients, but rather determine if a subset of patients with presumptive mitochondrial disease is more sensitive to sevoflurane than others. Also, the current study was not designed to demonstrate the superiority of sevoflurane over any other anesthetic protocol nor that hypersensitivity to sevoflurane indicates a complex I defect. As is reported in previous articles, we did find that a volatile anesthetic was well tolerated in this entire group of patients¹².

It is important to note limitations of this study. First, given the wide range of possible causes of mitochondrial disease, our sample size is small and certainly does not represent the potential breadth of presentations which could impact the concentrations of sevoflurane needed during induction. In addition, the small sample size seriously limits our ability to control for confounders. As such, these data should be considered preliminary and will require additional work to validate our findings. Second, as noted earlier, the maintenance phase of the cases in this study were not standardized, thus any variation in sensitivities between groups for this phase of an operation would be more difficult to demonstrate. Third, and perhaps most importantly for the occurrence of complications, the surgical cases were all rather short, seldom taking more than one hour to complete. We have noted over the past decade that mitochondrial patients do not tolerate prolonged cases as well as other patients. Finally, the technique we used to determine sensitivity does not represent steady state sensitivity. To do so would require a prolonged induction that would be difficult to support. Our results must be interpreted with each of these caveats in mind.

We think that a slow induction is in the best interest of mitochondrial patients in cases where anesthetic sensitivity is unknown. Our induction can be compared to slow inductions in other frail patients (such as the very elderly) where drug is slowly titrated to reach the desired effect, in this case a BIS of 60. The key feature of this protocol was a carefully titrated induction of general anesthesia that was accompanied by data from a processed EEG monitor. At our institutions, processed EEG monitoring of these patients is now routine practice for children with known complex I dysfunction, or those coming in for a diagnostic muscle biopsy.

Summary

Our preliminary data indicate that, during induction, patients with complex I dysfunction are markedly hypersensitive to sevoflurane. Further investigation is required to validate our findings. Other types of mitochondrial dysfunction do not as commonly lead to hypersensitivity to this drug. While our results do not establish safety of this approach for mitochondrial patients, all patients in this study, including those with severe mitochondrial disease, were safely anesthetized using a sevoflurane-based anesthetic. Hallmarks of our care for mitochondrial patients were a slow anesthetic induction with sevoflurane and careful monitoring of their processed EEG.

Supplementary Material

Supplemental Data File (.doc, .tif, .pdf, etc., Published Online Only)_1

Supplemental Table S2. Details of the anesthetic course of patients as numbered in Table 1. Sevoflurane (BIS=60) is the sevoflurane concentration needed to reach a BIS of 60 at induction of general anesthesia. Sevoflurane maximum is the greatest concentration of sevoflurane used during maintenance of general anesthesia for each case. Medications and regional analgesia denote the ancillary techniques used to supplement the sevoflurane during maintenance of general anesthesia. F-I block indicates ultrasound-guided fascia Iliaca block. If only medications are listed, no regional analgesia was given. HR and BP Max are the maximal heart rate and blood pressure during maintenance. The mitochondrial defect is the primary mitochondrial defect determined. If no ETC abnormality was noted but the patient was still diagnosed as a mitochondrial patient by other criteria, the defect was termed No ETC Abnormality (NEA). Patients who were determined not to have a mitochondrial diagnosis were termed normal.

Supplemental Data File (.doc, .tif, .pdf, etc., Published Online Only)_2

Supplemental Table S1. Baseline characteristics of 91 patients enrolled in study. Eleven patients were withdrawn prior to induction of anesthesia. Where possible, baseline serum glucose (GLU) and venous pH/base excess (BE) were determined after anesthetic induction but prior to IV fluids or surgery. Note that patient #88 was on a ketogenic diet. ND indicates not done.

Key Points.

• Question: Do patients with mitochondrial complex I dysfunction have increased sensitivity to sevoflurane?

• Findings: Patients with complex I defects had increased sensitivity to sevoflurane compared to those with normal mitochondrial function.

• Meaning: Patients with mitochondrial disease should be monitored closely for hypersensitivity to volatile anesthetics.

Glossary of Terms.

BIS

Bispectral Index

EEG

Electroencephalogram

ET

End Tidal

ETC

Electron Transport Chain