Life-threatening adverse events following therapeutic opioid administration in adults: Is pharmacogenetic analysis useful?

Abstract

BACKGROUND:

Systemic approaches are needed to understand how variations in the genes associated with opioid pharmacokinetics and response can be used to predict patient outcome. The application of pharmacogenetic analysis to two cases of life-threatening opioid-induced respiratory depression is presented. The usefulness of genotyping in the context of these cases is discussed.

METHODS:

A panel of 20 functional candidate polymorphisms in genes involved in the opioid biotransformation pathway (CYP2D6, UGT2B7, ABCB1, OPRM1, COMT) were genotyped in these two patients using commercially available genotyping assays.

RESULTS:

In case 1, the patient experienced adverse outcomes when administered codeine and morphine, but not hydromorphone. Genetic test results suggested that this differential response may be due to an inherent propensity to generate active metabolites from both codeine and morphine. These active metabolites are not generated with hydromorphone. In case 2, the patient experienced severe respiratory depression during postoperative recovery following standard doses of morphine. The patient was found to carry genetic variations that result in decreased morphine efflux transporter activity at the blood-brain barrier and increased sensitivity to opioids.

CONCLUSIONS:

Knowledge of the relative contribution of pharmacogenetic biomarkers and their influence on opioid response are continually evolving. Pharmacogenetic analysis, together with clinical history, has the potential to provide mechanistic insight into severe respiratory depressive events in patients who receive opioids at therapeutic doses.

The influence of pharmacogenetic variation on analgesia and anesthesia has long been known (1). Genetic polymorphisms have been described in drug-metabolizing enzymes (2–8), transporters (9,10) and receptors (11–16) involved in the opioid response. However, there is controversy related to the clinical utility of pharmacogenetic testing in the context of anesthesiology (17,18). This issue has been exacerbated by the complexity of pain as an outcome measure, the limited characterization of the opioid analgesic pathway and the inherent heterogeneity in study designs and patient cohorts. Moreover, the majority of pharmacogenetic studies in this area have investigated a single gene in a cause-and-effect approach. Aside from the notable exception of cyto-chrome P450 2D6 (CYP2D6), the clinical impact of these single pharmacogenetic markers has remained elusive in the context of pain management.

Given our current understanding of the multiple markers that may influence opioid action and response, a polygenic analytical approach may be useful to determine how genetic determinants interact to influence pain response and the occurrence of adverse outcomes. Therefore, we developed a panel of 20 functional candidate polymorphisms in five genes (CYP2D6, UGT2B7, ABCB1, OPRM1 and COMT) that have been characterized in the opioid pharmacokinetic and pharma-codynamic pathways (9). This panel was tested in two patients who experienced rare and serious opioid-induced respiratory depression during or following surgical procedures in Toronto, Ontario. Together with a multidisciplinary team comprising the fields of anesthesiology, pharmacology, genetics, toxicology and surgery, we explored the value of post hoc pharmacogenetic analysis in the context of these two patients. The clinical utility of a pharmacogenetic analytical approach in predicting and preventing future opioid-related adverse outcomes from occurring is also discussed.

CASE PRESENTATIONS

Case 1

A 37-year-old woman (height 1.5 m, weight 58 kg) of French Canadian descent, with a history of six previous pregnancies resulting in two live births and three miscarriages, underwent elective caesarean section (Table 1). The patient experienced a previous episode of respiratory depression with codeine (single 60 mg dose), but tolerated hydromorphone based on self-report. Before surgery, she was given 0.75% hyperbaric bupivacaine, morphine 0.1 mg and fentanyl 0.01 mg intrathecally. In the operating room the patient received ketorolac 30 mg intravenously and acetaminophen 1300 mg parenterally. In the postanesthesia care unit, the patient received morphine 1 mg intravenously and tolerated it well. Twelve hours later, the patient complained of pain in the ward and received morphine 2 mg subcutaneously. Three hours later, a second dose of morphine 2 mg was administered subcutaneously. Thirty minutes after the second dose of morphine, the patient was not rousable to voice and painful stimulus and had a respiratory rate of 4 breaths/min. Heart rate and oxygen saturation were normal but blood pressure was elevated (152/90 mmHg). Pupils were equal and 2+ reactive. The resuscitation team arrived at the scene and intravenous naloxone was administered (0.16 mg plus 0.12 mg, for a total of 0.28 mg). The patient recovered consciousness immediately (Table 1).

TABLE 1.

Timeline of clinical events and opioid administration

Timeline	Doses administered and events
Case 1: Caesarean section
Perioperative	Hyperbaric bupivacaine 0.75% spinal
	Morphine 0.1 mg intrathecal
	Fentanyl 0.01 mg intrathecal
Postoperative	Morphine 1 mg intravenous
In ward
12 h after surgery	Morphine 2 mg subcutaneous
15 h after surgery	Morphine 2 mg subcutaneous
15.5 h after surgery	Respiratory rate 4 breaths/min; pupils 2+ reactive
	Naloxone 0.28 mg
	Patient recovered consciousness immediately
Postoperative	Prescribed hydromorphone 0.5 mg to 1 mg orallyevery 6 h as required
	Patient used hydromorphone 2 mg over 10 h
	Discharged
Case 2: Surgical resection of a complex intra-abdominal tumour
Perioperative	Desflurane at MAC 0.6 to 0.7
	Bupivacaine 0.2% 10 mL and morphine 2 mg: epidural load
Intraoperative	Bupivacaine 0.2% and morphine 0.025 mg/mL morphine: epidural infusion rate of 5 mL/h for 4 h
Recovery room
5 h after surgery	Extubated, stable condition
	Respiratory rate 16 breaths/min but unresponsive
5.5 h after surgery	Blood pressure 75/40 mmHg
	Phenylephrine 0.5 mg intravenous over 15 min
	Normal saline bolus 500 mL
	Ephedrine 25 mg intramuscular
	Epidural infusion discontinued
8 h after surgery	Level of consciousness did not improve; respiratory rate 4 breaths/min to 8 breaths/min (pH 6.64, PCO2 286 mmHg, PO2 121 mmHg)
8.5 h after surgery	Reintubated, hyperventilated (pH 7.18, PCO2 52 mmHg, PO2 537 mmHg)
Postoperative	Maximum pain: 1 of 10
	Total hydromorphone 0.6 mg intravenous over five days
	Discharged

MAC Minimum alveolar concentration; PCO₂ Partial pressure of carbon dioxide; PO₂ Partial pressure of oxygen

The patient was subsequently continued on diclofenac 50 mg orally every 8 h and acetaminophen 1000 mg orally every 6 h. She was prescribed hydromorphone 0.5 mg to 1 mg orally every 6 h as needed and 0.2 mg to 0.4 mg orally every 3 h as needed, and administered 0.5 mg at 14:35 and 16:30, and 1 mg at 24:00. After that, no additional opioids were administered and pain was managed solely with diclofenac and acetaminophen.

The patient was retrospectively genotyped for polymorphisms associated with codeine/morphine metabolism and response (Table 2). The patient carried a CYP2D6*2/*2 genotype associated with extensive/‘normal’ CYP2D6 enzymatic activity; however, she had a homozygous mutation in UGT2B7 (802T/T), which has been previously associated with increased formation of the pharmacologically potent morphine 6-glucuronide metabolite from morphine in some (19,20) but not all (21) studies. Additionally, the patient was a homozygous carrier of a COMT haplotype ‘TCA’ at positions 389, 611 and 675, respectively. This haplotype has been associated with slightly decreased catechol-o-methyltransferase activity and potentially higher opioid sensitivity (22–26).

TABLE 2.

Summary of genetic variants and genotype results

Gene	Variant(s)	dbSNP ID	Case 1	Case 2
CYP2D6	*2, *3, *4, *5, *6, *9, *10, *17, *29, *41, and gene duplication ×n		*2/*2	n/a
UGT2B7	802C→T	rs7439366	T/T	C/T
ABCB1	1236C→T	rs1128503	C/C	T/T
ABCB1	2677G→T/A	rs2032582	G/A	T/T
ABCB1	3435C→T	rs1045642	C/T	T/T
OPRM1	118A→G	rs1799971	A/A	G/G
COMT	389C→T	rs4633	T/T	C/C
COMT	611C→G	rs4818	C/C	C/C
COMT	675G→A	rs4680	A/A	G/G

dbSNP ID Single nucleotide polymorphism database identification; n/a Not applicable

Case 2

A 64-year-old woman of Filipino descent undergoing surgical resection of a complex intra-abdominal tumour received an epidural for perioperative pain control (Table 1). She was administered desflurane as a general anesthetic at a minimum alveolar concentration of 0.6 to 0.7 and the epidural was loaded with 0.2% bupivacaine and morphine 2 mg; an epidural infusion of 0.2% bupivacaine and 0.025 mg/mL morphine was maintained intraoperatively at a rate of 5 mL/h for 4 h. With the exception of 0.1 mg intravenous fentanyl at the time of induction, no additional opioids were administered and the patient was extubated at a minimum alveolar concentration of 0.7 and brought to the recovery room in stable condition, breathing spontaneously at a rate of 16 breaths/min but unresponsive. Her blood pressure declined to 75/40 mmHg within the first 30 min in the recovery room, which responded to phenylephrine 0.5 mg intravenously over 15 min, a bolus of 500 mL normal saline and ephedrine 25 mg intramuscularly; the epidural infusion was discontinued at this point after 5 h. The patient remained in the recovery room for an additional 3 h and her level of consciousness did not improve; her blood pressure and oxygen saturation were well maintained but her respiratory rate was between 4 breaths/min and 8 breaths/min during this period. At 3 h, her blood pressure again declined to 70/40 mmHg and she was unresponsive to repeated doses of intravenous ephedrine and phenylephrine. An arterial blood gas was drawn on 3 L fraction of inspired oxygen and revealed a pH of 6.64, a partial pressure of carbon dioxide of 286 mmHg and a partial pressure of oxygen of 121 mmHg. The patient was reintubated and hyperventilated; a repeat arterial blood gas drawn 30 min later showed a pH of 7.18, partial pressure of carbon dioxide of 52 mmHg and partial pressure of oxygen of 537 mmHg. The patient was drowsy but rousable 5 h to 6 h after admission to postoperative recovery and was extubated successfully the next morning. During her five-day postoperative recovery period, the patient’s self-rated maximal pain was very low (1 of 10) and she was administered a total of 0.6 mg intravenous hydromorphone (equivalent to approximately 3 mg of morphine) for pain management during this time. She was well controlled on acetaminophen 1000 mg orally every 6 h as the sole analgesic and discharged in stable condition five days postoperatively.

The patient was subsequently genotyped for polymorphisms associated with morphine biotransformation and response (Table 2). The most remarkable finding was that the patient carried homozygous mutations in both ABCB1 and COMT that may have predisposed her to morphine-induced respiratory depression. The haplotype ‘TTT’ in ABCB1 (at 1236, 2677, 3435), which leads to substantially decreased P-glycoprotein expression and activity, has been associated with increased systemic morphine exposure (27) and morphine accumulation in the brain (28). Similarly, the COMT haplotype ‘CCG’ (at 389, 611, 675) has been associated with markedly reduced enzyme activity and increased sensitivity to opioids (22,25,26). However, the interpretation of this case was complicated by the presence of a polymorphism in the μ opioid receptor (OPRM1 118G/G) that has previously been associated with increased morphine dose requirements in some studies (10,23,29).

DISCUSSION

These two cases illustrate the potential utility of pharmacogenetic analysis in elucidating the mechanism of respiratory depression in otherwise unexplained cases. In the context of case 1, the incidence of anesthesia-related complications related to childbirth is 0.5% (30), which amounts to 1700 deliveries per year in Canada, assuming a birth rate of 340,000. Of these complications, only a small number (2%) are related specifically to drug-induced central nervous system depression (30). This low incidence of respiratory depression in pregnant patients is due to several factors: they have high levels of endorphins; the pro-gesterone stimulates the respiratory centre; they are young and healthy; and their extracellular space is increased (31). Pharmacogenetic analysis revealed that this patient had the propensity to generate active metabolites from both codeine (extensive CYP2D6 activity) and morphine (increased UGT2B7 activity). Both the CYP2D6 extensive metabolizer phenotype and the UGT2B7*2/*2 genotype occur commonly in the Caucasian population (approximately 80% and 30%, respectively), but are less prevalent in other ancestral populations (32,33). Moreover, the tolerance to hydromorphone did not suggest that the sensitivity demonstrated in this case was inherent to the μ opioid receptor itself. This is further corroborated by the genetic findings. Furthermore, the timing of respiratory depression in this case is typical of intrathecal opioids (13 h after intrathecal injection). Thus, it is believed that the overall mechanism of opioid toxicity was an interaction of intrathecal and systemic opioid exposure in a patient sensitive to morphine.

The patient in case 2 carried genotypes corresponding to increased exposure and overall sensitivity to morphine and hydromorphone. Substantially decreased P-glycoprotein efflux transporter activity at the blood-brain barrier, in combination with low catechol-o-methyltransferase activity associated with increased sensitivity of the μ opioid receptor system, may have predisposed the patient to the adverse outcome reported here. However, her μ opioid receptor genotype (118G/G) has been associated in the literature with an increased morphine dose requirement. It is likely a balance of scales: more opioids are crossing the blood-brain barrier; there is high opioid receptor density; but the binding affinity of opioids to the receptor is weaker and/or the receptor activity is lower compared with the wild-type μ receptor genotype.

Evidently, the systemic picture and our pharmacogenetic interpretation is not clear based on our limited understanding of how these different markers interact with one another to protect against or exacerbate adverse outcomes. In addition, little is known about each of the individual markers, their mechanisms of action and their clinical usefulness in the context of complex and diverse patients. With additional cases, we may be better able to evaluate the predictive value of this panel of candidate genes and determine whether it may be useful in preventing adverse events or in identifying patients who may be at risk of complications due to opioids. In conjunction, functional studies may help shed light on how these polymorphisms may modulate treatment and response with opioid medications.

In the current context, the post hoc application of the opioid phar-macogenetic panel was useful in providing a mechanistic insight into the severe respiratory depressive events observed in these patients. However, because it was not clear how this information could prevent future adverse events from occurring, the patient’s clinical history remains the most important piece of information by which to guide future analgesic decisionmaking.