Abstract
Morphine-6-glucuronide, the active metabolite of morphine, and to a lesser extent morphine itself are known to accumulate in patients with renal failure. A number of cases on non-lethal morphine toxicity in patients with renal impairment report high plasma concentrations of morphine-6-glucuronide, suggesting that this metabolite achieves sufficiently high brain concentrations to cause long-lasting respiratory depression, despite its poor central nervous system penetration. We report a lethal morphine intoxication in a 61-year-old man with sickle cell disease and renal impairment, and we measured concentrations of morphine and morphine-6-glucuronide in blood, brain and cerebrospinal fluid. There were no measurable concentrations of morphine-6-glucuronide in cerebrospinal fluid or brain tissue, despite high blood concentrations. In contrast, the relatively high morphine concentration in the brain suggests that morphine itself was responsible for the cardiorespiratory arrest in this patient. Given the fatal outcome, we recommend to avoid repeated or continuous morphine administration in renal failure.
Introduction
Sickle cell disease (SCD) is an autosomal disorder that is characterized by the production of abnormal hemoglobin, resulting in chronic hemolytic anemia, susceptibility to pneumococcal and other infections, pain, stroke and multiple-organ dysfunction. 1 The hallmark of SCD are acute pain crisis, which are recurrent episodes of severe pain due to local vaso-occlusion, tissue infarction and the release of inflammatory mediators. 2 Opioids are widely accepted as the treatment of choice for severe acute pain during vaso-occlusive crisis. 1 –5 When morphine is applied, subcutaneous (s.c.) and intravenous (i.v.) bolus injections often result in poor sustained pain control, and i.v. infusion may be preferred to titrate morphine upon the patients reported pain. 3,6,7
Morphine is metabolized into two major metabolites, morphine-3-glucuronide (M3G) and morphine 6-glucuronide (M6G). Both metabolites are excreted renally and substantially accumulate in patients with renal failure. 8 –10 In addition, the plasma clearance of morphine itself can be markedly reduced due to kidney failure. 11,12 M3G has no opioid activity, whereas µ-opioid activity of M6G is greater than that of morphine. 13,14 M6G exerts analgesic activity and intrathecally applied M6G was shown to induce respiratory depression. 15 Upon i.v. morphine administration, M6G does normally not induce central effects because it is rapidly excreted via the kidneys and crosses the blood-brain barrier considerably slower and to a smaller extent than morphine. 16,17 However, evidence accumulates that patients with kidney impairment build up brain levels of M6G that are sufficient to induce long-lasting respiratory depression. 17,18 We report a lethal morphine intoxication in a patient with SCD that illustrates the clinical significance of brain accumulation of morphine and possibly M6G due to renal impairment.
Case report
A 61-year-old man with double heterozygous (HbSC) SCD and alpha thalassemia type 1 (deletions of 2 of the 4 genes coding for the alpha chains of hemoglobin) was admitted to our hospital with severe bone-pain in his right upper leg and hip due to a sickle cell crisis. Comorbidities were type 2 diabetes, sarcoidosis and hypertension. In addition, the patient displayed renal impairment which was likely a complication of SCD and/or diabetes. Serum creatinine was 344 µmol/L and the creatinine clearance, estimated by Cockcroft-Gault equation, was 24.8 mL/min. To relieve the pain, 50 mg tramadol was given t.i.d., and every 4−8 hours 5 mg morphine was s.c. injected for 5 days. On day 6, the pain worsened and it was decided to apply morphine i.v. by continuous infusion (Figure 1). The initial infusion rate was 5 mg/h and in order to titrate the morphine upon the pain, an i.v. bolus of 5 mg was applied 30 minutes after the start of the infusion. Subsequently, the infusion rate was increased to 7 mg/h. Three hours after the start of the infusion, drowsiness was observed and the infusion rate was reduced to 3.5 mg/h. At that time, the pain was still severe and the breathing frequency normal. Because the drowsiness increased, the infusion was discontinued 90 minutes later. Nonetheless, drowsiness further increased and 150 minutes after the end of infusion, the patient displayed a cardiorespiratory arrest. Despite basic life support and administration of 2 × 400 µg naloxone, 3 mg of atropine and 6 mg of adrenaline, the patient died after 40 minutes of resuscitation.
Time-line of morphine administration, resuscitation and death. s.c., subcutaneous; i.v., intravenous.
The total amount of morphine that the patient had received s.c. during the first 5 days was approximately 100 mg. On day 6, 10 mg morphine was administered s.c. and 29 mg i.v. (Figure 1). Autopsy was performed, but no cardiac, cerebral or other morphological alterations were detected that could explain the cause of death. However, in the cortex, discrete signs of ischemia were observed, suggesting that the patient died from respiratory arrest due to hypoxia as a result of hypoventilation that was caused by morphine. For confirmation, concentrations of morphine, M3G and M6G were determined in serum samples, obtained on day 3 of hospitalization and during resuscitation, and in post-mortem blood, brain and cerebrospinal fluid (CSF; Table 1). Samples were analyzed by a validated liquid chromatography-tandem mass spectrometry method. 19 Furthermore, the Naranjo score was calculated to determine the likelihood of whether the adverse drug reaction (in this case lethal intoxication) was due to morphine rather than the result of other factors. 20 The Naranjo score was 6, which indicates that this lethal intoxication was probably caused by morphine and/or its metabolites. 20
Concentrations of morphine, M3G and M6Ga
Abbreviations: M3G: morphine-3-glucuronide, M6G: morphine 6-glucuronide.
a Resuscitation: 150 min after the infusion was stopped.
Discussion
We here report the case of a patient with severe acute pain during a SCD crisis who was treated with morphine and died, most likely due to substantial accumulation of morphine, due to renal failure, leading to respiratory depression and cardiac arrest.
There are several case reports on morphine intoxication in patients with renal impairment. 18,21 –25 In plasma of these patients, substantial accumulation of M3G and M6G was found, and because M6G is a potent µ-opioid receptor agonist that can cross the blood-brain barrier, 13,17 the long-lasting signs of opioid intoxication were attributed to accumulation of M6G in the brain. In contrast, plasma clearance of morphine itself is less affected by kidney failure than that of M6G and morphine readily crosses the blood-brain barrier. 11 It seemed thus less plausible that morphine accumulated in the brain of these patients and caused the long-lasting central depressant effects that were observed. 18,21 –25
In contrast to plasma levels of morphine and M6G in relation to toxicity, data on brain or CSF concentrations are limited. There is one study that reports concentrations of morphine, M3G and M6G over a 48 period in plasma and CSF of neurosurgical patients, who received short term i.v. infusion with morphine. 17 Interestingly, the mean dose of morphine that was infused in that study in 30 minutes was 27 ± 5.5 mg, which is very close to the dose of 29 mg that was i.v. infused in our patient in 270 min. Mean maximal plasma concentrations of morphine, M3G and M6G in these neurosurgical patients were approximately 300, 1000 and 250 ng/mL, respectively. 17 The serum concentrations in our patient, 150 minutes after the infusion was stopped, were 71.1, 1899 and 466 ng/mL, respectively. This comparison demonstrates that our patient indeed displayed substantial accumulation of M3G and M6G, especially since morphine in our patient was infused over a much longer time period. Moreover, serum was obtained 150 min after the infusion was stopped, suggesting that maximal serum concentrations of morphine and metabolites were even higher. In addition, the serum concentrations of M3G and M6G in the sample that was obtained on day 3 (after administration of approximately 20 mg/day s.c. morphine on 3 consecutive days) were already higher than one would expect based on pharmacokinetic data obtained in various clinical situations. 26 It is thus likely that accumulation during the first 5 days also contributed to the high levels of M3G and M6G that were measured in the sample that was obtained on day 6 during resuscitation. Furthermore, in the study on the neurosurgical patients, the mean CSF concentrations of morphine, M3G and M6G were 20, 17 and 4 ng/mL, respectively. 17 Our patient had substantial higher post-mortem concentrations in CSF of morphine (29.6 ng/mL) and M3G (67.5 ng/mL). And although we could not quantify M6G in CSF of our patient, we cannot exclude that M6G in concentrations below 10 ng/mL (limit of quantification) was present in the CSF and contributed to the respiratory depression. Especially because M6G was found to be a 7.8-fold more potent µ-opioid receptor agonist than morphine itself. 13 In addition, the post-mortem concentration of morphine in brain tissue of our patient was 50.9 ng/g. Interestingly, simulations of morphine brain concentrations in 12 healthy volunteers, in whom respiratory depression had been observed, indicated that brain concentrations were ~41 ng/g. 27 Furthermore, in vitro experiments showed that morphine, M3G and M6G were stable in blood and urine at 4, 18−22 and 37°C for 10 days, 28 indicating that in the brain of our patient, post-mortem conversion of M3G and M6G to morphine or degradation of morphine is not to be expected. It is thus likely that morphine itself was responsible for the cardiorespiratory arrest in our patients, although we cannot rule out the contribution of M6G.
For renally compromised patients, recommendations on morphine dose reductions are available, e.g. for a glomerular filtration rate (GFR) of 20−50 mL/min a reduction to 75% of the normal dose is recommended. 29 In clinical practice, it is not straightforward to interpret these recommendations, as morphine demand and tolerance vary largely among patients. Our patient had an estimated creatinine clearance of 24.8 mL/min and we carefully titrated the i.v. morphine infusion according to the patients' response. The total morphine dose that was administered s.c. on days 1−5 was approximately 100 mg, whereas on day 6 another 10 mg was given s.c. and 29 mg i.v. (see Figure 1). This total morphine dose is considerably lower compared to other studies using morphine for acute pain crisis in SCD. 3,7 However, despite the fact that our patient received a relatively low total dose of morphine, substantial accumulation of morphine, M3G and M6G occurred due to renal impairment (vide supra and Table 1).
This lethal morphine intoxication illustrates that repeated or continuous morphine should preferentially be avoided in patients with renal impairment. Nonetheless, when patients with renal impairment do need treatment with strong analgesics, opiates should be considered that do not accumulate (both parent drug and active metabolites) in case of kidney failure. From that perspective, methadone, buprenorphine and piritramide might be suitable alternatives, as the use of these analgesics appears to be relatively safe in patients with renal impairment. 10,18,25,29
Footnotes
This research received no specific grant from any funding agency in the public, commercial, or not for-profit sectors.