Lost to follow-up: Post-operative polysomnography in at-risk, pediatric obstructive sleep apnea.

OBJECTIVE: Analyze adherence to AASM recommendations for post-operative polysomnography in eligible pediatric patients. STUDY DESIGN: Retrospective Cohort. SETTING: Tertiary, Outpatient Sleep Lab. METHODS: We conducted a retrospective analysis of pediatric patients, ages 1-17, previously diagnosed with moderate-severe obstructive sleep apnea that completed a surgical intervention. Chart review included demographic data, a co-morbidity of interest, the presence of an otolaryngology, primary care, or sleep medicine encounter, time to follow-up, the presence of a post-operative polysomnography, time to post-operative polysomnography, and the presence of an annual follow-up with any provider. RESULTS: Of the 373 patients, 67 patients met inclusion criteria. Fifty-nine followed-up with any provider, with 21 completing post-operative polysomnography. Patients with residual or recurrent symptoms (p < 0.01) and all patients with severe obstructive sleep apnea (p = 0.04) were more likely to complete post-operative polysomnography (PSG). Sub-analysis across at-risk categories (isolated moderate, isolated severe, moderate & a co-morbidity, and severe & a co-morbidity) revealed patients with severe obstructive sleep apnea & a co-morbidity completed a follow-up PSG more often than isolated moderate obstructive sleep apnea (p = 0.01). There was a difference in follow-up with sleep medicine across at-risk categories (p < 0.01). CONCLUSION: Recurrent symptoms and increasing disease severity were associated with obtaining post-operative polysomnography. However, variability existed for which patients completed post-operative polysomnography. We speculate an inconsistent standard across disciplines, inadequate post-operative obstructive sleep apnea management education, and uncoordinated systemic processes contribute to this discrepancy. Our findings support a standardized, multi-disciplinary care pathway for the management of at-risk, pediatric obstructive sleep apnea.

Exercise capacity remains supernormal, though mildly reduced in middle-aged military personnel with Moderate to Severe Obstructive Sleep Apnea.

INTRODUCTION: The relationship between moderate to severe OSA and exercise capacity remains unclear. Prior studies showing a reduction in VO2 max in this population have mostly involved middle-aged, overweight patients. We aimed to study this relationship in a similarly aged population of military personnel with previously undiagnosed moderate to severe OSA. METHODS: We studied late-career male military personnel who underwent CPET and polysomnography (PSG). Patients were categorized either into an OSA group (apnea-hypopnea index (AHI) ≥ 15 events/h) or a control group (AHI < 15 events/h). VO2 max was compared between groups. RESULTS: 170 male military personnel met criteria for the study. Mean AHI was 29.0/h in the OSA group (n = 58) versus 7.4/h in the controls (n = 112) while SpO2 nadir was slightly lower (86.0% vs. 89.0%). Patients were of similar age (53.1 vs. 53.7 years), and BMI was slightly higher in the OSA group (27.5 kg/m2 vs. 26.3 kg/m2). Percent-predicted VO2 max was supernormal in both groups, though it was comparatively lower in the OSA group (117% vs. 125%; p < 0.001). CONCLUSIONS: Military personnel with moderate to severe OSA were able to achieve supernormal VO2 max values, yet had an 8% decrement in exercise capacity compared to controls. These findings suggest that OSA without significant hypoxemia may not significantly influence exercise capacity. It remains likely that the effects of untreated OSA on exercise capacity are complex and are affected by several variables including BMI, degree of associated hypoxemia, and regularity of exercise. Statistically lower VO2 max noted in this study may suggest that untreated OSA in less fit populations may lead to significant decrements in exercise capacity.

Use of the 8-F angio-seal vascular closure device in large-caliber arteriotomies.

PURPOSE: To evaluate the feasibility and safety of using the 8-F Angio-Seal vascular closure device (VCD) to seal large-caliber (>8-F) access sites during percutaneous endovascular interventions. METHODS: A retrospective review was undertaken of 42 consecutive patients (34 men; mean age 67.8 years, range 36-94) undergoing percutaneous peripheral interventions with sheaths ranging from 9-F to 12-F and subsequent closure using 8-F Angio-Seal VCDs. Single-wall puncture (n = 48) of the common femoral artery was guided by ultrasound in 46 cases and palpation in 2. Forty procedures required therapeutic heparinization during the interventional procedure; protamine was administered in only 5. Per protocol, manual pressure was held for 15 minutes. Clinical and/or imaging follow-up was available in all cases within 3 months after the procedure. RESULTS: Immediate technical success was achieved in all cases, with hemostasis obtained within 5 minutes (no oozing or hematoma). The overall complication rate was 4.1% (2/48); one hematoma requiring surgical repair occurred 10 hours after VCD deployment. An asymptomatic pseudoaneurysm was discovered on follow-up imaging and was treated with ultrasound-guided thrombin injection with complete resolution. CONCLUSION: The use of the 8-F Angio-Seal VCD to close large-caliber arteriotomies ranging from 9-F to 12-F is feasible and safe, with a low complication rate.

Retrograde migration and endovascular retrieval of a venous bullet embolus.

Venous bullet embolism is a rare and complicated occurrence reported in approximately 0.3% of penetrating trauma. The management of bullet emboli is decided on a case-by-case basis, balancing the risk of the embolus itself against those associated with extraction. We report a case of a 19-year-old man who sustained a gunshot wound to the anterior chest, which migrated to the left internal iliac vein in a retrograde fashion. We were able to successfully retrieve the missile using an endovascular approach, thereby minimizing the morbidity associated with an open procedure.

Impact of obesity on outcomes after open surgical and endovascular abdominal aortic aneurysm repair.

BACKGROUND: This study examined impact of obesity on outcomes after abdominal aortic aneurysm repair. STUDY DESIGN: Data were obtained from the Veterans Affairs National Surgical Quality Improvement Program. Body mass index (BMI) was categorized according to National Institutes of Health guidelines. Multivariate regression adjusted for 40 other risk factors to analyze trends in complications and death within 30 days. RESULTS: We identified 2,201 patients undergoing 1,185 open and 1,016 endovascular aneurysm repairs (EVAR) for abdominal aortic aneurysms from January 2004 through December 2005. BMI distribution was identical in both groups and reflected national population statistics: approximately 30% were normal (BMI 18.5 to 24.9), 40% were overweight (25.0 to 29.9), and 30% were obese class I (30.0 to 34.9), II (35.0 to 39.9), or III (>/=40.0). After open repair, obesity of any class was independently predictive of wound complications (adjusted odds ratio = 2.4; 95% CI, 1.5 to 5.3; p = 0.002). Class III obesity was also an independent predictor or renal complications (adjusted odds rato = 6.3; 95% CI, 2.2 to 18.0; p < 0.0001) and cardiac complications (adjusted odds ratio = 4.5; 95% CI, 1.1 to 22.9; p = 0.045. After EVAR, obesity (any class) was predictive of wound complications (adjusted odds ratio = 3.1; 95% CI, 1.1 to 8.1; p = 0.026), but not predictive of other complications or death. Between the two types of operation, there were fewer complications and deaths after EVAR compared with open repair across all BMI categories, but outcomes were most disparate among the obese. CONCLUSIONS: Obesity is an independent risk factor that surgeons should consider during patient selection and operative planning for abdominal aortic aneurysm repair. Obese patients appear to particularly benefit from successful EVAR over open repair, but if open repair is required, special attention should be paid to cardiac risk, perioperative renal protection, and aggresive wound infection prevention measures.

Quality of life of patients with Takayasu's arteritis.

OBJECTIVE: Takayasu's arteritis (TA) is a chronic immune vasculitis that causes inflammation of the aorta and its branches and is clinically characterized by exacerbations and remissions. This study examined the quality of life (QoL) of patients with TA using the Medical Outcomes Study Short Form 36 (SF-36) Health Survey, a validated health related QoL questionnaire. METHODS: Questionnaires that included the SF-36 and demographic related variables were mailed to 392 patients enrolled in the Takayasu's Arteritis Research Association. Raw SF-36 scores, as well as Physical Health Summary (PHS) and Mental Health Summary (MHS) scores, were calculated according to standard protocols. Data were analyzed for predictors of superior QoL using univariate and stepwise logistic regression analysis. SF-36 scores were also compared with those of other chronic diseases associated with peripheral vascular disease (PVD) published in the literature. Results are reported as mean +/- standard error of the mean. RESULTS: A total of 158 patients (144 women, 14 men) with average age of 42.2 +/- 1.1 years responded to the questionnaire. Mean onset of symptoms occurred at 30.5 +/- 1.2 years, with a mean age at diagnosis of 34.7 +/- 1.2, and a median of four doctors were seen before diagnosis. The group underwent 299 TA-related surgical procedures (1.9 +/- 0.3), including coronary (38%), carotid (35%), upper extremity (30%), and lower extremity (26%) revascularization. PHS and MHS summary scores (39.2 +/- 1.0 and 44.5 +/- 1.0, respectively) were worse than mean scores for an age-matched healthy population as well as nationally reported scores for diabetes mellitus, hypertension, and coronary artery disease (all P < .0001). Multivariate predictors of better physical QoL were younger age (P = .003) and remission of the disease (P = .0002). The use of immunomodulating medications was associated with inferior physical QoL (P = .02). The sole predictor of better mental QoL was remission of disease (P = .002). CONCLUSION: TA is a rare disease with profound consequences on QoL. Scores for physical and mental health are worse compared with many other chronic diseases associated with PVD. Superior physical QoL is seen in younger patients, whereas inferior physical QoL is encountered in those who take immunomodulating medications. Because the only factor to influence positively both physical and mental QoL is disease remission, every effort should be directed to attenuate disease activity.

Effect of folic acid and vitamins B6 and B12 on microcirculatory vasoreactivity in patients with hyperhomocysteinemia.

OBJECTIVE: Hyperhomocysteinemia (HHcy) has been identified as an independent risk factor for atherosclerotic vascular disease. The effect of high-dose folic acid or combination vitamin therapy for the treatment of HHcy on the microcirculation is unknown. The purpose of this study was to evaluate the effect of a combination of folic acid, vitamin B6, and vitamin B12 on endothelium-dependent and endothelium-independent vasoreactivity in patientswith HHcy. METHODS: Baseline cutaneous microvascular vasoreactivity was measured in 20 patients with HHcy and 18 patients with normohomocysteinemia (NHcy). Laser Doppler scan imaging before and after iontophoresis of 1% acetylcholine chloride (endothelium-dependent response) and 1% sodium nitroprusside (endothelium-independent response) was performed for the measurement of forearm skin vasodilatation. Patients were then treated with 10 mg folic acid, 100 mg vitamin B6, and 1 mg vitamin B12 orally once a day for 6 months. Follow-up fasting serum homocysteine and cutaneous Laser Doppler scan imaging before and after iontophoresis were performed at 1, 2, 3, and 6 months. Statistical analysis was performed using Fisher's exact test, paired t test, and Wilcoxon matched-pairs signed-ranks test, with significance set at P < .05. RESULTS: The HHcy group was older than the NHcy group (70.89 +/- 1.95 vs 61.78 +/- 2.73 years, P = .02). Otherwise the groups were similar in terms of race, tobacco use, comorbid diseases, and serum lipoproteins. Over the 6-month period, fasting serum homocysteine levels decreased significantly in both the NHcy group (10.40 +/- 0.59 micromol/L vs 8.97 +/- 0.84 micromol/L, P = .01) and the HHcy group (19.80 +/- 1.06 micromol/L vs 13.40 +/- 0.86 micromol/L, P = .0002). There were no statistically significant changes in endothelium-independent vasoreactivity (voltage change from baseline) in either group. Endothelium-independent vasore activity decreased over the 6-month period in the HHcy group (0.20 +/- 0.04 V vs 0.11 +/- 0.03 V, P = .03). Subanalysis of HHcy with diabetes or age greater than 65 years both showed worsening trends in endothelium-independent vasoreactivity (P = .05 for both groups). There were no statistically significant changes in endothelium-independent vasoreactivity in the NHcy group. CONCLUSIONS: High doses of folic acid and vitamins B6 and B12 lower fasting serum homocysteine levels in patients with HHcy. Older and diabetic patients with HHcy tend to do worse possibly because of long-term fixed microvascular insult secondary to multiple sustained comorbidities.

The effects of the type of anesthesia on outcomes of lower extremity infrainguinal bypass.

OBJECTIVE: Three main types of anesthesia are used for infrainguinal bypass: general endotracheal anesthesia (GETA), spinal anesthesia (SA), and epidural anesthesia (EA). We analyzed a large clinical database to determine whether the type of anesthesia had any effect on clinical outcomes in lower extremity bypass. METHODS: This study is an analysis of a prospectively collected database by the National Surgical Quality Improvement Program (NSQIP) of the Veterans Affairs Medical Centers. All patients from 1995 to 2003 in the NSQIP database who underwent infrainguinal arterial bypass were identified via Current Procedural Terminology codes. The 30-day morbidity and mortality outcomes for various types of anesthesia were compared by using univariate analysis and multivariate logistic regression to control for confounders. RESULTS: The NSQIP database identified 14,788 patients (GETA, 9757 patients; SA, 2848 patients; EA, 2183 patients) who underwent a lower extremity infrainguinal arterial bypass during the study period. Almost all patients (99%) were men, and the mean age was 65.8 years. The type of anesthesia significantly affected graft failure at 30 days. Compared with SA, the odds of graft failure were higher for GETA (odds ratio, 1.43; 95% confidence interval [CI], 1.16-1.77; P = .001). There was no statistically significant difference in 30-day graft failure between EA and SA. Regarding cardiac events, defined as postoperative myocardial infarction or cardiac arrest, patients with normal functional status (activities of daily living independence) and no history of congestive heart failure or stroke did worse with GETA than with SA (odds ratio, 1.8; 95% CI, 1.32-2.48; P < .0001). There was no statistically significant difference between EA and SA in the incidence of cardiac events. GETA, when compared with SA and EA, was associated with more cases of postoperative pneumonia (odds ratio: 2.2 [95% CI, 1.1-4.4; P = .034]. There was no significant difference between EA and SA with regard to postoperative pneumonia. Compared with SA, GETA was associated with an increased odds of returning to the operating room (odds ratio, 1.40; 95% CI, 1.20-1.64; P < .001), as was EA (odds ratio, 1.17; 95% CI, 1.05-1.31; P = .005). GETA was associated with a longer surgical length of stay on univariate analysis, but not after controlling for confounders. There was no significant difference in 30-day mortality among the three groups with univariate or multivariate analyses. CONCLUSIONS: Although GETA is the most common type of anesthesia used in infrainguinal bypasses, our results suggest that it is not the best strategy, because it is associated with significantly worse morbidity than regional techniques.

Hypoxia attenuates insulin-induced proliferation and migration of human diabetic infrapopliteal vascular smooth muscle cells.

The proliferative effects of insulin on infrapopliteal vascular smooth muscle cells (VSMCs) have been established. We examined the effect of hypoxia in the presence and absence of insulin on the proliferation and migration of human diabetic infrapopliteal VSMCs in vitro. VSMCs isolated from the infrapopliteal arteries of male diabetic patients of identical disease and clinical patterns undergoing below-knee amputation were harvested and grown to subconfluence. Cells were then exposed to control medium (M199/1% fetal bovine serum/2% antibiotic-antimycotic) or control medium with 100 ng/mL insulin in oxygen concentrations of 17% (normoxia), 5%, and 1%. Cellular proliferation was assayed using [methyl-3H]-thymidine incorporation. Migration assays were performed using the Corning Costar Transwell system. Lactate dehydrogenase was assayed and compared among groups as a marker for cytotoxicity. VSMCs in normoxic conditions (17%) had a significant increase in both proliferation (100 +/- 6.5% vs. 124 +/- 4.7%, p = 0.007) and migration [73.2 +/- 9.3 vs. 118.1 +/- 14.9 cells/4 high-power fields (HPF), p = 0.03] when exposed to insulin. Of cells exposed to insulin, those at both 5% (75.9 +/- 7.9%, p = 0.0001) and 1% (73.6 +/- 4%, p < 0.0001) hypoxia proliferated at a significantly decreased rate compared with cells at normoxia (124 +/- 4.7%). Migration of these insulin-exposed cells was significantly decreased at 1% hypoxia (63.1 +/- 9.0 cells/4HPF) compared to those at normoxia (118.1 +/- 14.9 cells/4HPF, p = 0.006) and 5% hypoxia (101.2 +/- 10.0 cells/4HPF, p = 0.01). There were no significant differences in migration between cells at normoxia and 5% hypoxia. Finally, hypoxia and insulin exerted no significant effect on cytotoxicity. The proliferative and promigratory effects of insulin on diabetic VSMCs are attenuated in hypoxic conditions in a manner unrelated to cytotoxicity.

Thiamine (vitamin B1) improves endothelium-dependent vasodilatation in the presence of hyperglycemia.

Brachial artery vasoactivity (BAVA) is a reliable, noninvasive method of assessing endothelium-dependent vasodilatation (EDV) in vivo. Acute hyperglycemia, impaired glucose tolerance (IGT), and diabetes mellitus impair EDV, a precursor to atherosclerosis. Thiamine is a coenzyme important in intracellular glucose metabolism. The purpose of this study was to evaluate the effect of thiamine on BAVA in the presence of hyperglycemia. Ten healthy subjects (group H, mean age 27 years), 10 patients with impaired glucose tolerance by World Health Organization criteria (group IGT, mean age 65 years), and 10 patients with non-insulin-dependent diabetes mellitus (group NIDDM, mean age 50 years) were studied. Duplex ultrasound was used to measure brachial artery flow changes in response to reactive hyperemia following brachial artery tourniquet occlusion for 5 min. This test was performed after a 10 hr fast and at 30, 60, and 120 min after a 75 g oral glucose challenge along with measurements of blood glucose level (BGL). A week later, BAVA evaluation was repeated after administration of 100 mg of intravenous thiamine. BAVA (% increased blood flow) at peak and trough BGL was compared with and without thiamine. BAVA at peak glucose improved from 69.0 +/- 6.4% to 152.8 +/- 22.9% in group H (p < 0.005), from 57.6 +/- 12.6% to 139.7 +/- 12.4% in group IGT (p < 0.005), and from 57.8 +/- 8.3% to 167.8 +/- 11.6% in group NIDDM (p < 0.005) following administration of thiamine. On the other hand, at trough glucose levels, BAVA remained essentially unchanged in group H (prethiamine 83.8 +/- 6.5% vs. post-thiamine 83.8 +/- 17.0%, p > 0.05) as well as group IGT (prethiamine 96.7 +/- 8.5% vs. post-thiamine 104.0 +/- 17.4%, p > 0.05). BAVA at trough glucose was not measured in group NIDDM secondary to trough BGL > 140 mg/dL. EDV was improved by thiamine in the presence of hyperglycemia in healthy subjects and in patients with IGT and NIDDM. The mechanism by which thiamine improves EDV is not due to a glucose-lowering effect as thiamine had no effect on EDV under normoglycemic conditions. Routine administration of thiamine might improve endothelial function and therefore slow the development and progression of atherosclerosis, especially in patients with IGT and NIDDM who are prone to develop accelerated atherosclerosis.

Carotid plaque morphologic characteristics.

Carotid artery atherosclerosis is predominantly believed to mirror atherosclerosis elsewhere in the body. Endothelial injury results in expression of cell surface adhesion molecules with expression of sequence of genes involved in the inflammatory pathway and expression of proinflammatory cytokines. The combination of the inflammatory mediators and contribution by monocytes infiltrating the intima and vascular smooth muscle cell proliferation result in the development of atheromatous plaque with a lipid-rich necrotic core. Complications of these atheromatous plaques can lead to plaque instability, rupture, and subsequent hemorrhage or ulceration. The significant risk factors, characteristics associated with symptoms, and available diagnostic imaging modalities are also discussed with review of the relevant literature.

Evaluation of macrocirculatory endothelium-dependent and endothelium-independent vasoreactivity in vascular disease.

Macrocirculatory endothelium-dependent and independent vasodilatation is integral to tissue-bed oxygen delivery and homeostasis. Dysfunction of macrocirculatory vasoreactivity is a precursor to atherosclerosis and occurs in a similar fashion in multiple tissue beds long before the onset of symptoms. Impaired macrocirculatory vasodilatation has been shown to occur in certain disease states including diabetes mellitus, hypercholesterolemia, chronic renal failure, peripheral arterial atherosclerosis, and abdominal aortic aneurysms, as well as secondary to smoking, advanced age, menopause, high-fat diet, and sedentary lifestyle. Brachial artery vasoreactivity is a noninvasive means of assessing macrocirculatory vasodilatory capacity that may help identify patients at increased risk for peripheral and cardiovascular disease and allow for objective assessment and monitoring of treatment. Endothelium-dependent vasoreactivity, or flow-mediated dilatation, is measured after brachial artery occlusion with a pneumatic blood pressure cuff, and endothelium-independent vasoreactivity is measured after the administration of sublingual nitroglycerin. The accuracy of brachial artery vasoreactivity is dependent on hematologic variables, as well as diurnal, day-to-day, ultrasound operator, and reader variations; however, the overall coefficient of variation is only 1.8%. We discuss the importance of the macrocirculation, investigative methods for evaluating macrocirculatory vasoreactivity, and review the literature of vasoreactivity in these different states.

Evaluation of the microcirculation in vascular disease.

Insufficient blood flow through end-resistance arteries leads to symptoms associated with peripheral vascular disease. This may be caused in part by poor macrocirculatory inflow or impaired microcirculatory function. Dysfunction of the microcirculation occurs in a similar fashion in multiple tissue beds long before the onset of atherosclerotic symptoms. Impaired microcirculatory vasodilatation has been shown to occur in certain disease states including peripheral vascular disease, diabetes mellitus, hypercholesterolemia, hypertension, chronic renal failure, abdominal aortic aneurysmal disease, and venous insufficiency, as well as in menopause, advanced age, and obesity. Microcirculatory structure and function can be evaluated with transcutaneous oxygen, pulp skin flow, iontophoresis, and capillaroscopy. We discuss the importance of the microcirculation, investigative methods for evaluating its function, and clinical applications and review the literature of the microcirculation in these different states.

Aortic angiosarcoma presenting as distal arterial embolization.

Aortic angiosarcoma is a rare, malignant neoplasm of the vasculature, with 24 case reports in the literature. Patients usually present with either aneurysmal or occlusive disease. Treatment consists of en bloc resection of the vessel and contiguous structures with postoperative chemoradiation therapy. Despite surgery and adjunctive measures, prognosis remains dismal as a result of early metastasis and late diagnosis. We report a case of aortic angiosarcoma that presented with distal thromboembolic phenomenon.

Strategies of arteriovenous dialysis access.

Surgical management of the patient who requires hemodialysis access, while continuing to demand more attention from the vascular surgeon, suffers from discrepancies of approach and strategy. With the increase in incidence of dialysis dependent renal failure among our population, many have attempted to present a uniform, logical strategy with which the vascular surgeon can most effectively treat the hemodialysis patient in the long term. Most notably, the multidisciplinary Dialysis Outcomes Quality Initiative (DOQI) guidelines present the surgeon with a rough outline of hemodialysis access insertion strategy, and it has become nationally recognized as an acceptable summary of treatment strategy and goals. The decision as to the most appropriate surgical access to offer a patient depends on immediate need for hemodialysis, history and physical examination findings, and suitability of available veins in the extremity. While percutaneous, catheter based access affords the luxury of immediate access, these devices suffer from several complicating factors, such as infection, and damage to large, proximal veins. For long-term access, the autogenous access, while perhaps less successful in the immediate short term, is always the preferred access type given its favorable longevity. The surgeons should focus on sites distally on the extremity, reserving proximal sites for potential future access insertions should the primary access fail. In the absence of suitable vein, prosthetic access may be considered. When both the upper and lower aspects of both upper extremities have been exhausted, the surgeon should consider access insertion elsewhere, such as the lower extremity.

Medical factors affecting patency of arteriovenous access.

Arteriovenous access failure is multifactorial in nature with contributions from both medical and surgical etiologies. Medical causes of arteriovenous access failure are rare, and therefore infrequently identified as a major contributing source of malfunction. Although they account for only 10-15% of all cases of access failure, their importance should not be underestimated, especially in cases where a surgical source cannot be identified. Most medical causes are derived from Virchow's triad of endothelial cell injury, stasis, and hypercoaguability. Endothelial cell injury occurs through oxidative stress, activated platelets, increased levels of circulating tumor necrosis factor-alpha, and preexisting intimal hyperplasia. Stasis can occur through prolonged access compression, hypotension, or hypoalbuminemia. Finally, patients with renal failure requiring hemodialysis are frequently at increased risk for hypercoaguable states, except for situations of platelet dysfunction, and therefore access failure. Potential treatments include identifying and removing the offending source, as well as innovative, new medications to prevent their reoccurrence. Treatment is aimed at improving quality of life, as well as decreasing morbidity and hospital admissions in this difficult patient population.