Heparin-induced thrombocytopenia thrombosis after cardiac surgery. A case report

We report a rare case of cerebral infarct resulting in brain death due to heparin-induced thrombocytopenia thrombosis (HITT), manifested in the immediate postoperative period following aortic valve replacement in a 46-year-old woman whose only prior exposure to heparin was during catheterization four months prior to surgery. The diagnosis of HITT was suspected after a significant decrease of the platelet count and it was confirmed by a heparin-induced platelet activation assay showing platelet aggregation in the presence of heparin.     

Heparin-induced thrombocytopenia and thromboembolism in the postoperative period

The heparin-induced thrombocytopenia and thrombosis (HITT) syndrome is associated with hemorrhage as well as development of systemic thrombosis. A case is presented in which a posthepatectomy patient had probable heparin-induced thrombocytopenia complicated by venous thrombosis and pulmonary emboli after receiving low doses of heparin as line flushes. HITT is reviewed and factors related to its successful management in this postoperative patient are analyzed in detail.     

Immediate vein graft thrombectomy for acute occlusion after coronary artery bypass grafting

A 76-year-old man underwent coronary bypass grafting 3 days after exposure to heparin. Immediately after chest closure, he developed acute graft thrombosis and cardiac arrest in the setting of thrombocytopenia. Immediate graft thrombectomies were performed. Postoperative tests for heparin-induced thrombocytopenia and thrombosis (HITT) were positive. This case represents a dramatic example of HITT after coronary revascularization.     

Heparin-induced thrombocytopenia with iliacofemoropopliteal thrombosis in a patient operated for colorectal carcinoma liver metastases

We report a case of heparin-induced thrombocytopenia thrombosis (HITT) syndrome in a patient prophylactically treated with low molecular weight heparin. A 66-year-old men underwent radiofrequency-assisted partial liver resection for colorectal carcinoma liver metastases a year-and-a-half after he had been operated for rectal cancer. In the postoperative period, patient was prophilactically treated with reviparin sodium. On the 8th postoperative day, the platelet count decreased by more than 50% without clinical signs of thrombosis. HITT syndrome was suspected on the 19th postoperative day, after iliacofemoropopliteal thrombosis had developed, and related diagnosis was supported by the strongly positive particle gel agglutination technique immunoassay. Heparin was withdrawn and alternative anticoagulant, danaparoid sodium, was introduced in therapeutic doses. Despite delayed recognition, favorable clinical outcome was achieved. HITT syndrome should be considered with priority among the possible causes of thrombocytopenia in a surgical patient on heparin.     

Off-pump right atrial thrombectomy for heparin-induced thrombocytopenia with thrombosis

This report describes a 72-year-old woman with atrial fibrillation who presented with lower extremity ischemia secondary to thromboembolism. After lower extremity thrombectomy, the patient developed heparin-induced thrombocytopenia with thrombosis (HITT). Her postoperative course was complicated by recurrent supraventricular and ventricular tachycardia, secondary to a mobile thrombus in the right atrium extending into the right ventricle. Because administration of heparin was contraindicated, the patient underwent off-pump right atrial thrombectomy during a brief period of inflow occlusion. Postoperatively, she was placed on lepirudin. Her platelet count normalized without any further thrombotic episodes, and she was discharged on warfarin.     

Heparin-induced thrombocytopenia after liver transplantation

BACKGROUND: Unfractionated heparin sodium (UFH) or low-molecular weight heparin (LMWH) is used in anticoagulant protocols at several institutions to prevent thrombosis after liver transplantation. Heparin-induced thrombocytopenia (HIT) is an adverse immune-mediated reaction to heparin, resulting in platelet count decreases of more than 50%. The frequencies of HIT after liver transplantation and platelet factor 4/heparin-reactive antibody (HIT antibody) positivity in liver transplantation patients, however, are unknown. PATIENTS AND METHODS: The 32 men and 20 women underwent living donor liver transplantation. We started LMWH (25 IU/kg/h) on postoperative day (POD) 1, switching to UFH (5000 U/d) on POD 2 or 3. The dose of UFH was changed according to the activated clotting time level. HIT antibody levels were measured the day before surgery and on POD 7 and 14. Platelet count was measured daily for 3 weeks. RESULTS: The average platelet counts preoperatively, and on POD 7, 14, and 21 were 65, 88, 149, and 169 x 10(9)/L, respectively. Two patients developed hepatic artery thrombosis on POD 11 and 19, respectively, although they were HIT antibody-negative and their platelet counts were stable. In 2 other patients, the platelet count decreased suddenly from 107 x 10(9)/L on POD 4 to 65 x 10(9)/L on POD 6 and from 76 x 10(9)/L on POD 7 to 33 x 10(9)/L on POD 9, respectively. The heparin-induced platelet aggregation test was negative in these patients. The percentage of HIT antibody-positive patients was 0.5% preoperatively, 5.6% on POD 7, and 5.6% on POD 14. None of the subjects/patients developed UFH-related HIT. CONCLUSIONS: In our series, the occurrence of HIT after liver transplantation was uncommon.     

Leukocytosis as an independent risk factor for cerebral vasospasm following aneurysmal subarachnoid hemorrhage

OBJECT: The identification of patients at an increased risk for cerebral vasospasm after subarachnoid hemorrhage (SAH) may allow for more aggressive treatment and improved patient outcomes. Note, however, that blood clot size on admission remains the only factor consistently demonstrated to increase the risk of cerebral vasospasm after SAH. The goal of this study was to assess whether clinical, radiographic, or serological variables could be used to identify patients at an increased risk for cerebral vasospasm. METHODS: A retrospective review was conducted in all patients with aneurysmal or spontaneous nonaneurysmal SAH who were admitted to the authors' institution between 1995 and 2001. Underlying vascular diseases (hypertension or chronic diabetes mellitus), Hunt and Hess and Fisher grades, patient age, aneurysm location, craniotomy compared with endovascular aneurysm stabilization, medications on admission, postoperative steroid agent use, and the occurrence of fever, hydrocephalus, or leukocytosis were assessed as predictors of vasospasm. Two hundred twenty-four patients were treated for SAH during the review period. One hundred one patients (45%) developed symptomatic vasospasm. Peak vasospasm occurred 5.8 +/- 3 days after SAH. There were four independent predictors of vasospasm: Fisher Grade 3 SAH (odds ratio [OR] 7.5, 95% confidence interval [CI] 3.5-15.8), peak serum leukocyte count (OR 1.09, 95% CI 1.02-1.16), rupture of a posterior cerebral artery (PCA) aneurysm (OR 0.05, 95% CI 0.01-0.41), and spontaneous nonaneurysmal SAH (OR 0.14, 95% CI 0.04-0.45). A serum leukocyte count greater than 15 x 10(9)/L was independently associated with a 3.3-fold increase in the likelihood of developing vasospasm (OR 3.33, 95% CI 1.74-6.38). CONCLUSIONS: During this 7-year period, spontaneous nonaneurysmal SAH and ruptured PCA aneurysms decreased the odds of developing vasospasm sevenfold and 20-fold, respectively. The presence of Fisher Grade 3 SAH on admission or a peak leukocyte count greater than 15 x 10(9)/L increased the odds of vasospasm sevenfold and threefold, respectively. Monitoring of the serum leukocyte count may allow for early diagnosis and treatment of vasospasm.     

Delayed onset of heparin induced thrombocytopenia--a case report]

A 64-year-old man underwent CABG. 18000 unit of bovine lung heparin was used during operation for extracorporeal circulation and small amount of heparin was injected until the 5th postoperative day to keep intravenous line patent. His postoperative course was uneventful until the 10th postoperative day when marked ecchymosis developed on his eyelid. Laboratory data showed marked thrombocytopenia (1.3 x 10(4)/microliters) and disseminated intravascular coagulation. The diagnosis of heparin-induced thrombocytopenia (HIT) was confirmed by platelet aggregation test with heparin. Gabexate mesilate and platelet-concentrates were administered and platelet count returned to normal within a week. HIT usually occurs during administration of heparin. In this case, it occurred 5 days after we stopped heparin. Delayed onset of HIT might be related to delayed metabolism of heparin due to postoperative fluid shift.     

Effects of a nitric oxide donor on and correlation of changes in cyclic nucleotide levels with experimental vasospasm

OBJECTIVE: Vasospasm after subarachnoid hemorrhage (SAH) may result from hemoglobin-mediated removal of nitric oxide (NO) from the arterial wall. We tested the ability of the long-acting, water-soluble, NO donor (Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-1,2-diolate (DETA/NO), delivered via continuous intracisternal infusion, to prevent vasospasm in a nonhuman primate model of SAH. METHODS: First, vasorelaxation in response to DETA/NO was characterized in vitro by using monkey basilar artery rings under isometric tension. Next, monkeys were randomized to undergo angiography, unilateral SAH, and no treatment (SAH only, n = 4) or treatment with DETA/NO (1 mmol/L, 12 ml/d, n = 4) or decomposed DETA/NO (at the same dose, n = 4). Vasospasm was assessed by angiography, which was performed on Day 0 and Day 7. Levels of cyclic adenosine monophosphate and cyclic guanosine monophosphate (cGMP) were measured in cerebral arteries on Day 7. RESULTS: DETA/NO produced significant relaxation of monkey arteries in vitro, which reached a maximum at concentrations of 10(-5) mol/L. In monkeys, angiography demonstrated significant vasospasm of the right intradural cerebral arteries in all three groups, with no significant difference in vasospasm among the groups (P > 0.05, analysis of variance). The ratios of cGMP or cyclic adenosine monophosphate levels in the right and left middle cerebral arteries were not different among the groups (P > 0.05, analysis of variance). There was no significant correlation between arterial cGMP contents and the severity of vasospasm. CONCLUSION: DETA/NO did not prevent vasospasm. There was no correlation between the severity of vasospasm and cyclic adenosine monophosphate and cGMP levels in the cerebral arteries. These results suggest that events downstream of cyclic nucleotides may be abnormal during vasospasm.     

The correlation between immunological reaction in the arterial wall and the time course of the development of cerebral vasospasm in a primate model

To investigate the role of immunological reactions in the development of cerebral vasospasm after subarachnoid hemorrhage (SAH), the authors studied the correlation between immune/inflammatory reactions in the arterial wall and the time course of vasospasm in primates. Twenty monkeys were divided into four groups of 5 animals each: 1) a control group of sham-operated animals, 2) animals subjected to angiography 3 days after the induction of SAH (3-day SAH), 3) animals subjected to angiography 7 days after SAH (1-week SAH), and 4) animals subjected to angiography 7 and 14 days after SAH (2-week SAH). To induce SAH, the main cerebral arteries on the right were dissected free of the arachnoid, and an autologous blood clot was placed around the arteries. To evaluate vasospasm, all animals underwent a baseline angiogram before SAH; angiography was repeated at different intervals in each group, as outlined above. Histopathological changes and the deposition of the immunoglobulin IgG in the arterial wall were evaluated immunohistochemically in each group. The cerebral arteries on the side of the clot showed evidence of mild vasospasm (-24.6% reduction) on the angiogram performed on Day 3, severe vasospasm (-51.7%) on Day 7, and mild vasospasm (-12.8%) on Day 14. The infiltration of inflammatory cells was most marked in the spastic arterial wall in the 1-week SAH group. In the 2-week SAH group, severe myonecrosis and intimal disruption were observed, even in the vessels that showed only mild vasospasm, and the inflammatory reactions had almost abated.(ABSTRACT TRUNCATED AT 250 WORDS)     

Heparin-induced thrombocytopenia and thrombosis

Heparin-induced thrombocytopenia (HIT) and heparin induced thrombocytopenia with thrombosis (HITT) ar rare complications associated with use of unfractionate heparin (UFH) or low-molecular-weight heparin (LMWH) HIT is a benign clinical condition characterized by a mil drop in platelet count with no clinical significance. HIT is an immune-mediated reaction associated with a wide spread "hypercoagulable" state resulting in arterial an venous thrombosis. There is a higher incidence of HIT with UFH use than with LMWH use. Orthopedic surger patients are at higher risk for developing HITT than are patients who receive prophylactic heparin for cardiovascular surgery or medical reasons. Therapy for patients suspected of having HITT should begin with immedi ate discontinuation of heparin in any form followed by pharmacologic inhibition with thrombin (e.g., recombinant hirudin [lepirudin], argatroban, danaparoid sodium).     

Heparin-induced thrombocytopenia in patients administered heparin solely for hemodialysis

BACKGROUND: Heparin, universally used in patients on dialysis, is the cause for immune-mediated heparin-induced thrombocytopenia (HIT). METHODS: From an HIT registry, six patients were identified who received recent heparin solely for dialysis, developed HIT, and were treated with argatroban. Platelets counts, aPTTs, argatroban dosing, and outcomes were assessed. RESULTS: Before HIT was diagnosed, unfractionated heparin was used in doses of 2,000-12,000 units. The mean platelet count fell from 122 +/- 62 x 10(9)/L to 35 +/- 22 x 10(9)/L, and one patient experienced thrombosis. After HIT was diagnosed, heparin was discontinued, and argatroban therapy (mean dose, 1.7 +/- 0.9 microg/kg/min) was administered for 6.5 +/- 4.5 days (mean aPTT, 66.1 +/- 12.4 s). Patients continued on renal replacement therapy. Platelet counts increased during argatroban therapy. A 37-day composite endpoint of death, amputation, or new thrombosis occurred in four (66.7%) patients: three patients died from causes unrelated to thrombosis, and one patient developed new thrombosis after argatroban was discontinued. One patient experienced a hematocrit drop during treatment without overt bleeding or a need for transfusion. CONCLUSIONS: HIT can occur in patients administered heparin solely for hemodialysis. When HIT is suspected, heparin should be discontinued and an alternative anticoagulation initiated. Argatroban, which is not renally cleared, supports continued renal replacement therapy in HIT patients.     

Heparin-induced thrombocytopenia in open heart surgical patients: sequelae of late recognition.

Most patients undergoing open heart operations have had exposure to heparin for diagnostic and/or therapeutic procedures. Heparin antibody formation and heparin-induced thrombocytopenia with repeat heparin administration can cause high morbidity and mortality from thrombotic complications, especially when delay in diagnosis occurs. From 1981 to 1991, heparin-induced thrombocytopenia was diagnosed in 82 of 4,261 open heart surgical patients (1.9%). Platelet counts less than 100 x 10(9)/L (100,000/microL) or new or recurring thrombotic events prompted suspicion of heparin-induced thrombocytopenia. Heparin-dependent antibody was diagnosed preoperatively in 12 patients (group I) and postoperatively in 70 patients (group II). Heparin was not given postoperatively in group I patients, and complications in this group were limited to bleeding in 3 patients. There were no thromboembolic events and all patients survived. Group II patients had late recognition of heparin-dependent antibody postoperatively, and heparin exposure was continued for varying periods postoperatively. Thirty-seven group II patients (53%) had bleeding complications and 31 (44%) had thromboembolic complications. These complications led to death in 23 group II patients (33%). Heparin-dependent antibody may occur in patients having open heart operations and is a major cause of morbidity and mortality if not diagnosed early with cessation of heparin therapy.     

Thrombocytopenia in renal failure in a developing country.

A retrospective study over a 3-year period was done looking at predialysis platelet levels, in particular, thrombocytopenia. Seventy-five patients with acute renal failure (ARF) and 75 patients with chronic renal failure (CRF), treated at King Edward VIII Hospital, were randomly chosen. Platelet counts were performed on a coulter counter (S + 2) and counts of less than 150 x 10(9)/L were considered as thrombocytopenia. Of the 75 CRF patients, 47 were males. Eleven (14.7%) were thrombocytopenic with a mean platelet count of 118.3 x 10(9)/L and a range of 83-146 x 10(9)/L. The mean creatinine level was 1510 micrograms/L. The remaining nonthrombocytopenic patients had a mean platelet count of 268 x 10(9)/L and a mean creatinine of 1080 micrograms/L. Of the ARF patients, 39 were males. Twenty-two (29.3%) had thrombocytopenia with a mean platelet count of 98 x 10(9)/L and a range of 22-147 x 10(9)/L. The mean creatinine level was 819 micrograms/L. The remaining nonthrombocytopenic patients had a mean platelet count of 319 x 10(9)/L and a mean creatinine of 1020 micrograms/L. In CRF patients no correlation was found between thrombocytopenia and the disease process. Creatinine levels appear to be higher in the thrombocytopenia group than in the nonthrombocytopenic group. In the ARF group of patients, females had a higher frequency of thrombocytopenia than males. Obstetrical and gynecological causes and herbal ingestion were the 2 major underlying etiologies in the thrombocytopenic group. Thrombocytopenia appears to be a common presenting feature in ARF as opposed to CRF, and this may be accounted for by the underlying etiologies in ARF.     

Intraoperative management of patients with heparin-induced thrombocytopenia.

For 11 patients with confirmed heparin-induced thrombocytopenia, we used reversible platelet inhibition with iloprost, a stable prostacyclin analogue, to permit safe heparin administration for cardiac (n = 9) or vascular (n = 2) operations. In vitro, iloprost (0.01 mumol/L) prevented both heparin-induced platelet aggregation and 14C-serotonin release in all patients. Therefore, intraoperatively, a continuous infusion of iloprost was started before administration of heparin and was continued until 15 minutes after administration of protamine. For cardiac patients, after heparin administration, the whole blood platelet count did not change (171,000 +/- 29,000/microL versus 174,000 +/- 29,000/microL, mean +/- standard error of the mean); no spontaneous platelet aggregation was observed, and plasma levels of the alpha-granule constituents platelet factor 4 and beta-thromboglobulin increased from 38 +/- 14 and 140 +/- 18 ng/mL to 591 +/- 135 and 235 +/- 48 ng/mL, respectively. Fibrinopeptide A levels actually decreased from 287 +/- 150 to 27 +/- 6 ng/mL. Furthermore, adenosine diphosphate-induced platelet activation was preserved, postoperative bleeding times were unchanged, and no heparin-related deaths occurred. Similar results were obtained in both vascular patients. We conclude that temporary platelet inhibition with iloprost now permits safe heparin administration in all patients with heparin-induced thrombocytopenia who require a cardiac or vascular operation.     

Association between cerebrospinal fluid levels of asymmetric dimethyl-L-arginine, an endogenous inhibitor of endothelial nitric oxide synthase, and cerebral vasospasm in a primate model of subarachnoid hemorrhage

OBJECT: Decreased availability of nitric oxide (NO) has been proposed to evoke delayed cerebral vasospasm after sub-arachnoid hemorrhage (SAH). Asymmetric dimethyl-L-arginine (ADMA) inhibits endothelial NO synthase (eNOS) and, therefore, may be responsible for decreased NO availability in cases of cerebral vasospasm. The goal of this study was to determine whether ADMA levels are associated with cerebral vasospasm in a primate model of SAH. METHODS: Twenty-two cynomolgus monkeys (six control animals and 16 with SAH) were used in this study. The levels of ADMA, L-arginine, L-citrulline, nitrites, and nitrates in cerebrospinal fluid (CSF) and serum were determined on Days 0, 7, 14, and 21 following onset of SAH. Cerebral arteriography was performed to assess the degree of vasospasm. Western blot analyses of the right and left middle cerebral arteries (MCAs) were performed to assess the expression of eNOS, type I protein-arginine methyl transferase (PRMT1) and dimethylarginine dimethylaminohydrolase (DDAH2). Cerebrospinal fluid levels of ADMA remained unchanged in the control group (six animals) and in animals with SAH that did not have vasospasm (five animals; p = 0.17), but the levels increased in animals with vasospasm (11 animals) on Day 7 post-SAH (p < 0.01) and decreased on Days 14 through 21 (p < 0.05). Cerebrospinal fluid levels of ADMA correlated directly with the degree of vasospasm (correlation coefficient = 0.7, p = 0.0001; 95% confidence interval: 0.43-0.83). Levels of nitrite and nitrate as well as those of L-citrulline in CSF were decreased in animals with vasospasm. Furthermore, DDAH2 expression was attenuated in the right spastic MCA on Day 7 post-SAH, whereas eNOS and PRMT1 expression remained unchanged. CONCLUSIONS: Changes in the CSF levels of ADMA are associated with the development and resolution of vasospasm found on arteriograms after SAH. The results indicate that endogenous inhibition of eNOS by ADMA may be involved in the development of delayed cerebral vasospasm. Inhibition of ADMA production may provide a new therapeutic approach for cerebral vasospasm after SAH.     

Is vasospasm related to proliferative arteriopathy?

Although proliferative arteriopathy has been postulated to play a role in the etiology of vasospasm after subarachnoid hemorrhage (SAH), histological and morphological studies examining cerebral vasospasm have produced conflicting results. To help settle this controversy, the authors used an in vivo label of cell division, bromodeoxycytidine, to assess cell proliferation in a primate model of SAH. Fifteen cynomolgus monkeys received a clot of either whole blood (11 animals) or red blood cells (four animals) placed around the right middle cerebral artery (MCA). On the day of surgery continuous intravenous infusion of bromodeoxycytidine was begun and continued until the animal was sacrificed immediately after arteriography on Day 7, 12, or 27 following surgery. Sections from the right and left MCA's were stained with a monoclonal antibody against bromodeoxcytidine, and labeled cells were counted. Arteriographic evidence of vasospasm occurred in nine monkeys on Day 7. On Day 12 and Day 27 no monkeys had persistent vasospasm. Placement of subarachnoid clot around the right MCA increased proliferative activity across all layers of the arterial wall. Most of the labeled cells were in the adventitia and the endothelium. Although there were more dividing cells in all layers of the right MCA than the left MCA (p < 0.01), the number of stained cells per section was limited (range 0.1 to 21.2, mean 8) and the occurrence of vasospasm was not associated with the number of dividing cells in the right MCA on Day 7, 12, 27, or for all days combined (p > 0.6). Cerebral vasospasm after SAH was not associated with the extent of proliferation of cells in the vessel wall, nor could the intensity of the limited proliferative changes have been responsible for narrowing of the vessel diameter.     

Etiology of cerebral vasospasm in primates

A primate model was used to determine whether oxyhemoglobin (OxyHb), methemoglobin (MetHb), or bilirubin is likely to be responsible for cerebral vasospasm following subarachnoid hemorrhage (SAH). Forty cynomolgus monkeys were randomly assigned to one of five groups. On Day 0, each animal underwent angiography followed by right craniectomy and placement of an Ommaya reservoir with its catheter adjacent to the right middle cerebral artery (MCA). The animals received intrathecal injections twice a day for 6 days of one of the following solutions: mock cerebrospinal fluid (CSF); OxyHb; MetHb; bilirubin; or supernatant fluid from an incubated mixture of autologous blood and mock CSF. On Day 7, angiography was repeated and the animals were killed. Comparison of angiograms obtained on Day 0 and Day 7 of the experiment showed significant vasospasm of the right MCA and the right anterior cerebral and internal carotid arteries in the animal groups that had received OxyHb or supernatant fluid. There was a smaller reduction in diameter of the same vessels in the bilirubin group (not statistically significant), while no effects were observed in the groups receiving MetHb or mock CSF. Electron microscopy of the right MCA's gave results consistent with the angiographic findings. One monkey in the OxyHb group developed a delayed-onset right MCA infarction. These data suggest that OxyHb is the cause of cerebral vasospasm following SAH.     

Neuropeptide Y in the primate model of subarachnoid hemorrhage.

The cause of cerebral vasospasm after subarachnoid hemorrhage (SAH) remains unknown. Recently, an association between the potent vasoconstricting peptide, neuropeptide Y, and delayed cerebral vasospasm after SAH has been postulated. This was based on the findings of increased neuropeptide Y levels in the cerebrospinal fluid (CSF) and plasma after SAH in animals and humans. For this study, the primate model of SAH was used to assess the possible role of neuropeptide Y in delayed vasospasm after SAH. Fifteen cynomolgus monkeys underwent placement of a clot of either whole blood or red blood cells in the subarachnoid space around the middle cerebral artery (MCA). Sequential arteriography for assessment of MCA diameter and sampling of blood and CSF for neuropeptide Y were performed: before SAH (Day 0); 7 days after SAH, when signs of delayed cerebral vasospasm peak in this model and in humans; 12 days after SAH; and 28 days after SAH. Subarachnoid hemorrhage did not evoke changes in CSF or plasma levels of neuropeptide Y. Nine monkeys had arteriographic evidence of vasospasm on Day 7, but no change in neuropeptide Y levels occurred in plasma or CSF. In addition, neuropeptide Y levels did not change, even after resolution of vasospasm on Day 12 or Day 28. Neuropeptide Y levels were substantially higher in CSF than in arterial plasma (p less than 0.003 at each interval). No correlation was found between neuropeptide Y levels in CSF and in plasma. These results do not confirm a relationship between neuropeptide Y levels in the CSF or peripheral plasma and delayed cerebral vasospasm in SAH.