Somatostatin Does Not Prevent Serotonin Release and Flushing during Chemoembolization of Carcinoid Liver Metastases

A 48-yr-old woman (163 cm, 54 kg) with carcinoid syndrome due to nonresectable liver metastases (two bulky metastases in the right lobe and diffuse small metastases in the left lobe) from an ileum carcinoid resected 2 yr previously was scheduled for a series of chemoembolizations via  the hepatic artery during monitored anesthesia care. With indium-111-octreotide scintigraphy (150 MBq), bulky nuclide enhancements were detectable in the right liver lobe and in the central liver, demonstrating an octreotide receptor–positive carcinoid tumor.

Except for flushing of the face three or four times daily, the patient was free of symptoms for 1.5 yr after her first operation. The symptoms worsened (flushing and diarrhea) 6 months ago, and the serum 5-HT concentration (1,430 μg/l; reference range, 110–330 μg/l) and urinary excretion of 5-hydroxyindole acetic acid (265 mg/24 h; reference range, 2–9 mg/day) were markedly increased. Therefore, octreotide acetate treatment was started and diminished the tumor-related symptoms, especially diarrhea. However, to achieve better and longer-lasting symptom control, chemoembolization of the bulky metastases was planned.

On the day of the preanesthetic examination the patient presented in slightly reduced physical condition and reported a history of intermittent flushing of the face, hypotension during flushing, and diarrhea. Findings of chest radiography, echocardiography, and laboratory tests were normal except for an increased serum 5-HT concentration (583 μg/l), an increased 5-hydroxyindole acetic acid urinary excretion (83 mg/day), and a slight increase in γ-glutamyl transferase (26 U/l; reference range, 4–18 U/l). The patient's medication for symptom control (200 μg octreotide acetate three times daily subcutaneously) was continued until the morning of chemoembolization.

For pressure measurements, arterial and central venous catheters were inserted via  the right internal jugular vein and left radial artery during local anesthesia before each chemoembolization, and the electrocardiogram (lead II) was monitored. Central venous blood was withdrawn for measurements of 5-HT concentrations in both whole blood and plasma. For measurements in whole blood, blood was withdrawn in an EDTA tube, immediately cooled in crushed ice, and stored at −20°C until analysis. For measurements in plasma, sampling tubes with EGTA and glutathione were used and centrifuged, and the plasma was stored at −20°C until analysis.

Prior to chemoembolization, the patient received prednisolone (250 mg intravenously), as well as the histamine H1 and H2 receptor antagonists clemastine (120 mg intravenously) and cimetidine (400 mg intravenously). Furthermore, somatostatin was infused continuously (0.1 μg · kg⁻¹· min⁻¹), as recommended for prophylaxis of carcinoid crisis. The patient was then brought to the angiography suite.

For pain control, the patient received the long-acting opioid piritramide (7.5 mg intravenously), metamizol (2.5 g intravenously), and a remifentanil infusion adjusted to deliver 40–120 μg/h as needed. 5-HT concentrations were measured during specified stages of the procedure (e.g. , at catheter placement in the hepatic artery and at embolization) and particularly during flushing.

Despite somatostatin infusion, hemodynamic conditions were quite unstable, and the patient showed severe flushing several times during the first session of chemoembolization of the bulky metastases in the right liver lobe. To treat hypotension, phenylephrine and theo-adrenalin–caffedrine were injected. Furthermore, 2.5–5 mg ketanserin, a specific 5-HT²-receptor antagonist, was repeatedly injected to control hypertension.

5-HT concentrations in whole blood were markedly increased even before intervention, but, despite flushing episodes, they varied little throughout the procedure (fig. 1). In contrast, 5-HT concentrations in plasma showed an almost sixfold fluctuation, between 205 and 1,578 μg/l (fig. 1). However, no apparent correlation was observed between 5-HT concentrations in plasma and either flushing episodes or hypotension.

The same prophylaxis was administered before the subsequent two chemoembolizations. While hemodynamic conditions during the second intervention (repeated embolization of the metastases in the right lobe 6 weeks later) were much more stable than in the first procedure, 5-HT concentrations in whole blood were also much increased at baseline and varied in a similar range (982–1,250 μg/l). Again, plasma 5-HT concentrations varied widely (85–1,072 μg/l). Although vasoconstrictor therapy was not necessary, ketanserin was used to control hypertensive episodes. Again, no apparent correlation was seen between 5-HT plasma concentrations and blood pressure alterations.

Before the third chemoembolization 4 months later, the patient's condition had much improved, octreotide had been withdrawn, flushing episodes were rare, and the diarrhea had stopped. Embolization of a small metastasis in the left liver lobe was hemodynamically uneventful. Initial 5-HT concentrations in whole blood, however, were even higher (1,352–1,640 μg/l) than at the previous baseline before interventions, but they did not change much during the procedure. Although 5-HT concentrations in plasma increased immediately from 240 to 330 μg/l after the start of embolization to 677 and 558 μg/l, no hypotension or bronchospasm occurred, despite the nausea reported by the patient.

Since patients with carcinoid syndrome undergoing tumor manipulation are prone to develop carcinoid crisis, avoidance of mediator release is an important goal.

Somatostatin and its longer-acting analog, octreotide, are commonly used for symptom control in patients with carcinoid syndrome. 8Perioperative administration of somatostatin or octreotide is recommended as intravenous prophylaxis for carcinoid crisis, either at a dosage of 100 μg/h or as a 50-μg single dose, in addition to preoperative administration of corticosteroids and antihistaminic drugs. 5,8Although the cellular mechanism of the action of somatostatin is still unclear, five subtypes of somatostatin G-protein-coupled receptors (sstr 1–5) have been identified. Most midgut carcinoid tumors express at least one somatostatin receptor subtype, and many express all five subtypes. 9Octreotide may bind to any of the five subtypes, but there is evidence that its action regarding the prevention of mediator release is strongly associated with sstr-2.

Among 25 carcinoid patients, all 7 patients with a positive octreotide scan but without expression of sstr-2 mRNA failed to respond to octreotide treatment. 10A missing sstr-2 is therefore probably responsible for about 20% of cases involving a positive octreotide scan without successful response to octreotide treatment. 10Polyclonal antibodies against sstr-2 for immunohistochemical studies, together with octreotide scintigraphy, may help to identify patients suitable for somatostatin analog treatment. 11 

However, our patient's octreotide scintigraphy results were positive, and good initial symptom control was achieved after the beginning of octreotide treatment, most likely indicating a sstr-2–positive tumor. Nonetheless, somatostatin failed to prevent 5-HT release during chemoembolization. Despite somatostatin infusion at an even higher dosage (0.1 μg · kg⁻¹· min⁻¹or 330 μg/h) than recommended, high baseline 5-HT concentrations in whole blood, increasing slightly from the first to the third procedure despite chemoembolization, failed to decrease. Furthermore, somatostatin failed to suppress the increase in 5-HT plasma concentrations occurring during each chemoembolization (i.e. , the 5-HT release from the tumor).

Different mechanisms of mediator release may explain the failure of somatostatin during chemoembolization of carcinoid tumors. Physiologic mechanisms of 5-HT release from enterochromaffin cells and carcinoid tumors include exocytosis of secretory granules, in which 5-HT and other mediators are stored. This process is influenced by a complex pattern of receptor-mediated mechanisms, including the inhibition of 5-HT release by somatostatin. 12 

During chemoembolization of carcinoid tumors, other mechanisms are likely involved, such as tissue hypoxia with loss of functional cell membrane integrity or mechanically induced cell damage, leading to a massive and fast mediator release and not preventable by somatostatin. On the other hand, recent investigations suggest that mechanically induced 5-HT release from carcinoid tumor cells is mediated via  a G-protein–coupled mechanoreceptor, and 5-HT release so evoked can be inhibited by somatostatin in a concentration-dependent manner. 13 

Despite marked increases in plasma 5-HT concentrations, clinical signs (e.g. , flushing and hemodynamic instability) did not correlate with 5-HT increases during chemoembolization. 5-HT release may mediate diarrhea, hypertension, and bronchoconstriction but not flushing and hypotension. 8Substance P and tachykinins, in turn, may be important for flushing, but their release is only partially suppressed by somatostatin. 14 

The observed discrepancy between 5-HT plasma and whole blood concentrations during chemoembolization is interesting and is most likely due to the fact that secreted serotonin is actively absorbed from plasma and mainly stored in platelets. 15The increase in baseline 5-HT whole blood concentration during the course of therapy might be caused by tumor necrosis following each embolization, with ensuing release of 5-HT, which is then stored in platelets. Accordingly, evaluation of rapid concentration changes as relevant to vascular receptors should be performed in plasma.

Different kinetics of 5-HT in whole blood and plasma and no apparent correlation between 5-HT concentrations and flushing or hypotension were also suggested by measurements in patients with carcinoid tumors undergoing surgery during neurolept anesthesia. 16However, we were able to demonstrate a relative reduction in the fluctuation of 5-HT concentrations in plasma during the chemoembolizations from the first to the third procedure, indicating a reduction of active tumor mass, but we were not able to show an absolute decrease in the basal concentrations before each procedure. Accordingly, the absolute plasma 5-HT concentration as an indicator of tumor mass reduction is doubtful.

To our knowledge, this is the first report of simultaneous 5-HT concentration measurements in plasma and whole blood in a patient with carcinoid syndrome undergoing somatostatin prophylaxis for chemoembolization. Despite prophylactic medication with somatostatin infusion, corticosteroids, and H¹- and H²-histamine receptor antagonist treatment, hemodynamic instability and flushing developed, and increased 5-HT plasma concentrations indicated substantial 5-HT release. Although 5-HT release was still detectable, the second and third chemoembolizations were less eventful hemodynamically, possibly because of tumor mass reduction evoked by the first embolization.

In contrast to our findings, 4-day pretreatment with octreotide, 100–200 μg subcutaneously daily, caused a significant decrease in baseline 5-HT whole blood concentrations and decreased their increase after pentagastrin stimulation. 175-HT concentrations in whole blood showed only minor variations during the succeeding chemoembolization in four patients with intraoperative prophylaxis (octreotide, 100 μg subcutaneously 4 times daily, and ketanserin, 20 mg intravenously 2 times daily), consistent with our findings, and hemodynamic variables were influenced little. However, the more sensitive 5-HT plasma concentrations were not measured. 17 

Fatal mediator release despite somatostatin prophylaxis is also suggested by a lethal carcinoid crisis, despite prophylactic administration of somatostatin (500 μg subcutaneously preoperatively and intravenously intraoperatively) in a patient undergoing tricuspid valve replacement. However, this patient did not receive steroids and antihistaminic drugs preoperatively, and 5-HT concentrations were not measured. 18 

In conclusion, prophylactic perioperative somatostatin infusion in a recommended dose of 0.1 μg · kg⁻¹· min⁻¹did not prevent serotonin release during chemoembolization of octreotide receptor–positive carcinoid metastases, despite successful preoperative octreotide treatment. Accordingly, sudden hemodynamic instability due to massive mediator release should be expected in patients with carcinoid syndrome, despite the use of currently recommended prophylactic medications.