In Depth Description

Reprinted from:

European Journal of Cardio-Thoracic Surgery Oxford Journals
Medicine & Health 
European Journal Cardio-Thoracic Surgery
Volume 29,  Issue 5
Pp.  748-759.


Patients with variant angina represent a diagnostic and therapeutic dilemma. Variant angina is a disease with various causes, variations in treatment guidelines and variable prognosis. In an era of robotic cardiac surgery and automatic cardioverter defibrillators, it is regrettable that we lack enough information on the optimal management of this entity. Lack of randomised trials and a tendency for spontaneous remissions makes interpretation of results difficult. We review the pathophysiology, presentation and recent developments in medical and surgical management of variant angina. Literature is full of conflicting data and it is difficult to make specific recommendations.

1 Introduction

Angina occurring spontaneously (angina at rest) associated with electrocardiographic changes was recognised as early as 1930s [1,2]. Prinzmetal et al. [3] first described this variant of angina pectoris as a distinct entity in 1959. They considered it to be caused by atherosclerosis affecting a single major coronary artery with intermittent ‘increased tonus’ [3,4]. MacAlpin and colleagues based on the angiographic findings in 12 patients of variant angina (VA) supported this concept [5,6]. Natural history of variant angina is not fully understood. It could lead to myocardial infarction (MI), life-threatening ventricular arrhythmias and sudden death [7–22]. Role of spasm in its pathophysiology was highlighted by many authors but it remained controversial [5,8–12,20–23]. The advent of angiography provided more insight into the pathophysiology.

2 Incidence (see also Section 4.1)

It is difficult to know the precise incidence of coronary artery spasm as different authorities have used different criteria to define it [24]. Besides, the incidence depends on the patient population studied (racial variation) and the provocative tests used [24]. A Japanese study found an incidence of 32.3% using intracoronary acetylcholine [25] and 29% using ergonovine [26] in patients undergoing coronary angiography for evaluation of chest pain. Similar studies in France [27] and North America [28] found a much lower incidence of 12.3% and 4%, respectively, of inducible coronary artery spasm using ergonovine test.


3 Mechanism

3.1 Anatomical status of coronary arteries
Coronary arteries could be normal (no anatomical narrowing) in a significant proportion of patients [29–37]. Reported incidence of normal coronary arteries varies from 30% [38] to 64% [39,40].

3.2 Abnormal vasomotor tone (spasm of the coronary arteries)
With the advent of coronary angiography, the occurrence of spasm moved from the realm of speculation and was documented conclusively [29–31,41–50]. The previously reported incidence of spasm in patients of variant angina (20–80%) [11,51], went up (70–95%) with the introduction of provocative tests [38].

Atherosclerotic disease affecting large coronary arteries altered their vasomotor tone and reactivity [9,31,52–55]. There was an intimate association of spasm with sites of organic stenosis. MacAlpin [56] found that in 88% cases spasm causing ischemia was localized to the site of an organic lesion. At angiography, identifying spasm in an already obstructed coronary artery could be difficult leading to underestimation of its occurrence [39,40].

Studies in animals demonstrated that partially obstructed coronary arteries undergo cyclical increase in luminal narrowing [57]. This cyclical phenomenon could be prevented by administration of vasodilators and aspirin. The initiating event could be spasm [8,58] or platelet aggregation [57]. During rest and sleep the diameter of the coronary arteries was smallest and tone was highest, a situation conducive to development of spasm; during exercise the tendency to spasm was minimised by dilatation of the artery and neurohumoral influences [5]. This could explain angina occurring at rest and the circadian variation present in most patients [59,60].

A variety of physiologic manoeuvres and pharmacologic agents could reproduce coronary spasm in patients of VA [61]. Thus, a diffuse, not localized, abnormality of the vasomotor tone was probably responsible [62]. An intrinsic abnormality in spontaneous tone generation and recovery could be involved as spasm could occur in the absence of spasmogens [63]. Vitamin C deficiency [64] and abnormalities of the endothelium (e.g. nitric oxide deficiency) [65–68] could contribute to vasomotor dysfunction.

It should be recognised that coronary artery spasm was not specific to Prinzmetal angina [69]. On selective coronary angiography, 0.5–0.8% of patients could show evidence of coronary artery spasm stimulated either by the tip of the catheter or the contrast medium [69–71].

3.3 Coronary emboli
Coronary thromboemboli, without causing significant obstruction, could induce coronary artery spasm and cause acute MI [72]. Such thromboemboli could occur in patients with prosthetic valves [73,74].

4 Predisposing factors

4.1 Racial predisposition
Compared to Caucasian patients, the Japanese patients have diffusely hyper-reactive coronary arteries with a high incidence of multivessel spasm [75]. Both genetic and environmental factors have been suggested to play a role [76–78]. Amongst genetic factors, missense gene mutation and amino acid substitution of the endothelial nitric oxide synthetase gene (resulting in deficient nitric oxide production) have evoked most interest [75–78]. In Japanese patients with MI Okumura et al. [79] and Fukai et al. [80] reported 69% and 75%, respectively, incidence of coronary spasm. Similarly, Pristipino et al. [81] found a much higher incidence of coronary spasm in Japanese patients of myocardial infarction as compared to Caucasians (64% vs 17%). The reported incidence of coronary spasm in Caucasian patients varies from 0% to 30% [48,82–86]. In France the incidence was 21% [82] while in Italy the incidence was 11% [83].

4.2 Age and sex
Patients with coronary spasm-induced cardiac arrest and those having angina with normal coronaries were generally younger than those with other causes of cardiac arrest [21,35].

A male predominance [13,17,18] was reported in coronary spasm-induced cardiac arrest while Chevalier et al. [21] found female predominance. A female predominance was found in variant angina with normal coronary arteries [34,87].

4.3 Cigarette smoking
In vasospastic angina, coronary endothelium becomes more sensitive to the deleterious effects of cigarette smoking [21,65].

Cigarette smoking adversely affected nitric oxide-mediated regulation of coronary artery tone [65]. Several reports suggested a strong association between cigarette smoking and pure coronary spasm, particularly in young women [88–90]. Myerburg et al. [15], on the other hand, did not find any such association.

4.4 Insulin resistance/hyperinsulinemia
Hyperinsulinemia secondary to insulin resistance was identified as risk factor for vasospastic angina [91–93]. Impaired glucose tolerance with late hypersecretion of insulin may contribute to the pathogenesis of severe coronary vasospasm [94].

4.5 Role of steroid hormones
Vascular hyper-reactivity could be a critical factor involved in the increased incidence of coronary artery vasospasm and ischemic heart disease in postmenopausal women [95–98]. Transdermal progesterone patch could prevent coronary vasospasm [99]. Prevention of vasospasm was via direct actions on coronary artery smooth muscle cells where it led to a reduction in the number of estrogen receptors [98]. However, there was some evidence to suggest that medroxyprogesterone, in contrast to progesterone, increased the risk of coronary vasospasm [100].

4.6 Impact of classical risk factors for coronary artery disease
Traditionally described risk factors for development of coronary artery disease such as total cholesterol, low-density lipoprotein, high-density lipoprotein, triglycerides, diabetes mellitus, and body mass index, etc. may not contribute to coronary vasospasm [89,90]. Only cigarette smoking was consistently associated with coronary artery spasm [88–90].

4.7 Others
Withdrawal from chronic industrial nitroglycerin exposure could lead to sudden intense coronary artery spasms, attacks of angina and sudden death [101].

Psychobehavioral patterns could influence coronary vasospasm. Alexithymia (deficient psychological awareness) was found to be twice as prevalent in patients with coronary spasm (31%) as compared to the control group (14%) [102].

5 Diagnoses

5.1 Classical symptoms
Previously, classical features described were recurrent angina at rest associated with »2 mm elevation of ST segment (in areas other than old infarcts); with normalisation of the ST segment after relief of pain without degenerating into myocardial infarction (serial ECG/myocardial enzymes) [38]. Other characteristic features included a tendency to recur cyclically and history of spontaneous remissions [6,103]. A circadian variation was found in most patients [59,60]. Recently, some authors emphasised that the manifestations of variant angina ranged from mild chest discomfort to myocardial infarction, life-threatening ventricular arrhythmias, sudden cardiac arrest and death [7–22].

5.2 Myocardial infarction
Myocardial infarction could occur in those suffering from isolated coronary artery spasm in spite of apparent clinical improvement with administration of calcium antagonist drugs [12,104,105]. It was more frequent and unpredictable in variant angina of recent onset [104,105]. Recurrent MI could occur without underlying coronary atherosclerosis [10,11]. The relatively small size of infarcts suggested that coronary reperfusion occurred in the early stages of myocardial infarction [80].

5.3 Sudden cardiac arrest and sudden death
Cardiac arrest and sudden death were important risks of variant angina even in the absence of severe organic coronary stenosis [7–22]. In fact, MacAlpin [22] found that the risk of one of these events in variant angina was increased 1.5-fold by the absence of high-grade coronary stenosis and three-fold by history of angina-linked syncopal attacks. Sudden death could occur due to coronary spasm without pain [9,54]. The risk of sudden death was much higher in patients with multivessel spasm and serious arrhythmias [106,107].

Ventricular fibrillation, tachycardia, and complete atrioventricular block could occur during ischemic episodes caused by spasm with or without anginal pain [107]. In variant angina, arrhythmias occurred in about 25% of patients with a tendency to recur in the same patient [106,107]. Recurrent transient complete heart block could occur leading to Stokes Adams attacks [108].

Some researchers are of the opinion that coronary spasm was the most frequent cause of cardiac arrest in cardiac arrest survivors with no apparent heart disease [13]. Recommendations have been made to perform a provocative test (ergonovine test) in all survivors of cardiac arrest if no underlying cardiac disease was found [13,21]. However, other researchers have found a low prevalence of coronary artery spasm in these patients [109].

5.4 Other distinguishing features
Patients of vasospastic angina with normal coronaries were usually young females with no typical signs of exertional angina [34,39,40]. Incidence of exertional angina was 11–36% in patients with normal coronaries and 70–75% in patients with fixed coronary artery stenosis [34,39,40].

It should be emphasised that there were no reliable clinical or electrocardiographic features to distinguish patients with normal coronaries [39,40]. Presence or absence of severe coronary artery obstruction could not be predicted from the severity of angina, presence of effort angina or the occurrence of a positive exercise test [39,40].

6 Investigations

6.1 Electrocardiography
Ischemic changes were predominantly in inferior leads in patients with isolated coronary spasm [34,39,40]. A majority of patients (90%) with normal coronaries had inferior ischemia compared to only 33% incidence of inferior ischemia in patients with anatomical narrowing [39,40]. The incidence of arrhythmias (premature ventricular beats, AV blocks, ventricular tachycardia) could be as high as 40% [5]. Although suggestions have been made regarding correlation of coronary arterial spasm with lead distribution of ST segment elevation, it has not received wide acceptance [56].

6.2 Angiography
The coronary arteries could be normal (isolated coronary artery spasm) or there could be associated narrowing of the arteries. One should be careful not to interpret catheter-induced dissection and/or spasm for coronary spasm [110,111]. This distinction is important to avoid unnecessary surgeries [110–112].

6.3 Provocative tests
In susceptible patients, coronary artery spasms could be induced by a variety of pharmacologic agents and even by some physiologic manoeuvres highlighting the non-specific hyper-reactivity of coronary vasculature in these patients [24,61].

6.3.1 Ergonovine test

The technique was described in detail by several authors [27,82,113–115]. It consisted of an intravenous injection of an ergot alkaloid (ergonovine maleate) which could be repeated if necessary [114]. Demonstration of spasm of the coronary vessels associated with chest pain and ECG changes was suggestive of presence of variant angina.

6.3.2 Sequential spasm provocation tests

In patients with rest angina, the incidence of provoked coronary spasm was relatively low with the standard single spasm provocation test. Sequential provocation tests (intracoronary injection of acetylcholine following an ergonovine test) could be more useful [115]. A triple sequential spasm test consisted of an initial acetylcholine test and then an ergonovine test and then, if no spasm was provoked, an intracoronary injection of acetylcholine [115]. It was proposed as potentially the most sensitive test [24].

6.3.3 Other provocative tests

Many chemicals such as serotonin [116], histamine [117] and dopamine [118] could induce coronary vasospasm. Their role as diagnostic agents was currently not well defined.

Exercise [119], cold pressor tests [120–122] and hyperventilation [123–126] have also been used to provoke coronary spasm. In patients with active variant angina, the sensitivity of the cold pressor and exercise tests were found to be too low [121–123]. In patients who become asymptomatic, either with treatment or spontaneously, the sensitivity of all tests decreased markedly [121–123]. Sensitivity of hyperventilation or provoked alkalosis was higher, but not as high as ergonovine [123–128]. Combined test (exercise and hyperventilation) has the advantage of being non-invasive with a high specificity but low sensitivity [24,125,126]. Kaski et al. [61] reported that spasm was provoked by ergonovine in 96% of patients, by hyperventilation in 54%, by histamine in 47%, by exercise in 46%, and by the cold pressor and handgrip tests in 11% and 7%, respectively. No significant differences were found in the responses to provocative tests of patients with normal coronary arteries or non-significant stenoses and those with significant lesions [61].

Clinical value of ergonovine test was its high sensitivity in patients with active variant angina and its predictable production of spasm which could be predictably and reliably reversed [27,117,120–122]. Nonetheless, deaths have been reported following ergonovine administration and for this reason the test was not universally accepted [104,121,122]. Cold pressor test has been advocated as a useful, low-risk method to assess efficacy of cardiac plexectomy [128].

The role of provocative tests in the clinical diagnosis of coronary spasm was controversial. This was reflected by variations in the clinical use of these tests between specialist cardiac centres. In North America and Europe, provocative tests were on decline, while in Asia it was still popular presumably because of the higher prevalence of symptomatic coronary artery spasm [24].

6.3.4 Ergonovine stress echo

It could identify regional wall motion abnormalities and was suggested as feasible, accurate and safe for the diagnosis of vasospastic angina [129–136]. It could diagnose coronary vasospasm otherwise missed by conventional non-invasive stress tests and even by coronary angiography [134,135]. It could replace invasive angiographic tests [133–136]. As yet, there are no long-term follow-up studies showing the benefits and clinical utility [134,135].

6.4 Radionuclide scintigraphy
Abnormal fatty acid metabolism was more often observed in ischemic myocardium than wall motion abnormalities in patients with vasospastic angina. Radiolabeled fatty acid analogues such as 123I-ß-methyl-p-iodophenyl pentadecanoic acid (BMIPP) and 123I-metaiodobenzylguanidine (MIBG), have been proposed as a tracer for detecting myocardial damage. Studies suggested that BMIPP and MIBG scintigraphy and MIBG single-photon emission computed tomography (SPECT) could be accurate non-invasive techniques for determining the presence and location of vasospasm [137–140]. Clinical utility of these tests need to be evaluated by further studies.

6.5 Hyperventilation echo
Hyperventilation performed under echocardiographic monitoring could be a safe screening test to unmask vasospastic myocardial ischaemia in patients with angina at rest, in whom documentation of spontaneous episodes is not available [141].

6.6 Intracoronary Doppler study
In patients with variant angina, characteristic serial changes in coronary flow velocity occur before occlusive spasm. These dynamic changes in coronary flow velocity could be monitored using a Doppler guidewire. Variant angina could be diagnosed earlier by monitoring flow velocity rather than by monitoring for ischemic electrocardiographic changes [142].

7 Management

Stopping smoking, as in all other forms of angina, was an important complementary measure [21]. In all other respects, the treatment of variant angina was different from typical angina pectoris [38]. Long-term management depended, to a large extent, on whether the spasm was occurring de novo in normal coronary arteries or whether it was superimposed on organic obstructive lesions [143].

7.1 Medical management

7.1.1 Beta blockers

Beta blockers which were so effective in management of typical angina pectoris did not relieve pain and in fact could precipitate and intensify pain in variant angina [30,32,49,144–147]. Nonetheless, some patients have been treated with beta blockers and reported to have an event-free clinical course [15,148–151].

7.1.2 Calcium antagonists and nitrates

Calcium antagonists could prevent contraction of smooth muscle fibres and relieve spasm. Several reports confirmed their clinical benefits, and therefore, they were the most commonly used drugs [38,152–162].

Compared to nitrates, nifedipine was more effective with fewer uncomfortable side effects and was preferred by the majority of patients [157–158]. Treatment with calcium antagonists decreased complication rate and improved survival without infarction as compared with other medical treatments [119,120]. Calcium antagonists should be continued even after bypass grafting in patients with vasospastic angina to prevent spasm in the artery distal to the graft [159]. Newer calcium channel blockers like benidipine might have better prognostic effects [163].

In many patients with variant angina, symptoms might disappear spontaneously and the ergonovine test could revert to negative [164]. Calcium antagonists could probably be safely discontinued in some patients; ergonovine test could be helpful in identifying such patients. Longer follow-ups to confirm that symptoms do not recur would be required [164].

In survivors of cardiac arrest with no underlying heart disease and demonstrable spasm of coronary arteries, calcium antagonists could prevent recurrence [13,15,17,18]. Some authors have reported a recurrence-free follow-up of 36 ± 18 months in this subset of patients [15]. However, there were reports of recurrence of cardiac arrest in spite of being on a combination of nitrates and calcium channel blocker [22].

7.1.3 Others

Amiodarone [161], ketanserin [165], prazosin [166,167], denopamine (adrenergic beta-1 agonist) [168], angiotensin converting enzyme inhibitors [169] and nicorandil [170,171] have been tested with variable clinical response. They have been proposed either as monotherapy or as an adjunct to calcium antagonists.

Increased activity of Rho-kinase could cause hypercontraction of vascular smooth muscles and was implicated in divergent cardiovascular diseases including vasospastic angina [172]. Recently, fasudil (intracoronary/intravenous infusion/per oral), a selective Rho-kinase inhibitor, was proposed as a novel therapeutic strategy for this group of patients [172–176].

Some studies have highlighted the limitations of medical therapy, including the administration of long-acting calcium antagonists [177]. It was effective in only 38% patients with pure coronary spastic angina in Japan [177]. Medical treatment showed a good response in female patients (63% vs 31% in males) and those with ST-segment elevation during selective spasm provocation tests [177]. Patients with a long history of chest pain attacks before hospital admission and those with diffuse spasms had poor responses [177].

7.1.4 Implantable cardioverter defibrillator (ICD)

ICD in addition to medical therapy could be useful in patients with life-threatening ventricular arrhythmias and recurrent sudden cardiac arrest secondary to coronary vasospasm [178–181].

7.1.5 Percutaneous interventions

Percutaneous intervention techniques represent an alternative treatment modality for patients with vasospastic angina refractory to aggressive medical therapy [182–184]. Stent placement has been reported to result in successful vasodilation that completely prevented anginal attacks [182,185]. There was no difference concerning the results of coronary intervention between the patients with variant angina and those with non-variant angina [186]. Risk of percutaneous intervention techniques was similar to that for other indications [186].

Kishida et al. [187] compared the three treatment modalities, i.e. calcium antagonists, nicorandil and nitroglycerin infusion and percutaneous transluminal coronary angioplasty (PTCA), respectively. The cumulative cardiac event rate was 22% at 1 year and 23% at 3 years for the first treatment, 11% at the same intervals in the second treatment and 6% at the same intervals in the third treatment modality [187]. Thus, apart from use of coronary vasodilators, early revascularisation should be attempted.

However, some authors have reported that following coronary angioplasty in patients with variant angina with organic lesions, symptoms due to coronary spasm usually persisted or recurred, often with restenosis [188,189]. Besides, spasm could occur at a site different from the initial stenosis, even in the absence of restenosis. Stents could be best advocated as an adjunct to medical therapy in the management of focal coronary artery spasm [190]. Calcium antagonists should be continued in patients with coronary spastic angina after stent placement [191].

Brachytherapy was recently proposed as treatment for in-stent restenosis and was the subject of clinical trials and debates [192]. Role of drug-eluting stents in coronary spasm was not defined at present [24].

7.2 Surgery

7.2.1 Coronary artery bypass grafting

Reported outcome of coronary artery bypass grafting (CABG) have ranged from disappointing [12,30,42,43,193] to gratifying [3,39,43,55,143,194]. These wide variations in results were partly because of the wide spectrum of angiographic findings in this disease ranging from nearly normal to severely stenosed vessels [41].

Some authors have suggested that these patients might be ideal candidates for surgery [193–195]. On the other hand, some authors would reserve surgery only for those with incapacitating symptoms refractory to medical therapy [6,43]. In between these two extreme opinions, there are those who advocated CABG for patients with significant organic obstructive disease only [41].

Surgical therapy provided relief but the results were still suboptimal as compared to other forms of angina. Review of literature suggested an operative mortality of 12–15%, postoperative MI in 24%, graft occlusion in 30–60% and recurrent chest pain in 30–78% of patients [39,43,196,197]. Clinical benefit is reported in less than 50% patients [38,43]. Isolated coronary artery spasm (no angiographically proven anatomical narrowing)
If surgery was performed in those with spasm but no anatomical narrowing, competitive flow would result in early graft occlusion [39]. Aortocoronary bypass grafting was therefore not recommended in patients with non-occluded coronary arteries [144,198]. Some investigators would not recommend bypass grafting in such patients even if they had severe disabling symptoms [43,199–200]. Possible reasons suggested for suboptimal outcome were a diffuse spasm occurring distal to the graft site with or without involvement of the graft anastomosis site [38,143,198]. Spasm could occur on an angiographically normal vessel which was not grafted [198].

Ono et al. [55], on the other hand, have reported extremely good clinical results with a pedicled in situ internal mammary artery graft in two patients with intractable variant angina refractory to medical management. Both patients had isolated coronary artery spasm on angiography. Patients were asymptomatic after more than 5 years of follow-up. Similarly, Sussman et al. [143] performed bypass grafting in two patients with coronary spasm and «20% atherosclerotic luminal narrowing with good clinical outcome. However, both Ono et al. [55] and Sussman et al. [143] have emphasised that bypass grafting in absence of anatomical narrowing should be reserved only for patients with life-threatening ischemic symptoms refractory to medical management. Associated anatomical narrowing of the coronary arteries
If the lesions were not operable, then medical management (nitrates and/or calcium channel blockers) was appropriate. If, on the other hand, the lesions were amenable to bypass grafting, the appropriate management was controversial [194–196]. In general, a consensus seem to be emerging that patients with high-grade obstructive lesions did well with CABG while those without significant obstructive lesions tend to do poorly [33,39].

Surgical success was related directly to the degree of coronary artery stenosis [39]. Pasternak et al. [40] found that in patients with significant obstructive lesions, 73% had marked improvement in anginal symptoms and the mortality rate of surgery was comparable to that with any other unstable angina group. It is of note that authors who reported favourable surgical outcome had patients with at least one fixed lesion with »70% stenosis [194–196,198,199–201]. Incidence of recurrent chest pain was 30% in presence of significant narrowing (»90%) of at least one vessel compared to 78% incidence if narrowing was not significant [39]. Schick et al. [199], in one of the largest series reported, found that if comparable clinical categories (stable, unstable and postinfarction angina) were made, then results of surgical revascularisation in patients with variant angina and fixed coronary artery disease were no different from that of classic angina. Focal versus diffuse spasm
Results of surgical revascularisation were likely to be more gratifying in patients with focal spasm [143,55]. In patients with demonstrable diffuse spasm of the coronary arteries, medical management would be more appropriate particularly in the absence of significant organic lesions [55]. On pump versus off pump
Some authors favoured the use of cardiopulmonary bypass as there was always a danger of inducing coronary spasm during the anastomotic procedure, which could affect the entire length of the native coronary artery [55]. Saphenous vein or internal mammary artery
Emerging evidence suggests that internal mammary artery might be a better conduit as compared to saphenous vein [55,201]. Internal mammary artery was known to have autoregulatory properties to adapt blood supply according to the demands of the myocardium. Besides, studies have demonstrated that ergonovine changed the calibre of saphenous vein graft significantly but it failed to do so with the internal mammary artery [202,203]. Saphenous vein grafts were prone to spasms late after bypass surgery [204,205]. Ligation of the proximal coronary artery
Competitive flow from the native vessel could lead to early graft occlusion [12]. To avoid this, some authors have described ligation of the proximal coronary artery, particularly in association with saphenous vein grafts [143]. Sussman et al. [143] suggested that absence of coronary ligation could allow vasoactive substances to initiate focal spasm which could get propagated along the whole length of the artery. However, some authors would advise caution in ligating the proximal coronary artery in view of the characteristic frequent spontaneous remissions of variant angina [55]. Techniques to prevent coronary spasm in perioperative period
Several additional measures were suggested to decrease the incidence of coronary spasm during and immediately after coronary revascularisation like addition of diltiazem in the cardioplegic solution, continuous intravenous infusion of nitroglycerin, an indwelling pigtail catheter in Valsalva sinus for bolus dosages of isosorbide dinitrate and periodic administration of nifedepine through the nasogastric tube [206]. Excision of the stenotic area
In cases with severe localised stenosis, insertion of an interposition graft could be combined with excision of the stenotic area [5]. The benefits obtained were questionable. Role of intra-aortic balloon pump (IABP)
Use of IABP in patients with intractable variant angina was associated with improved results [39,40,198,205]. IABP might augment the diastolic coronary artery perfusion, alter cadiovascular reflex tone and ameliorate the effects of spasm, decreasing significantly the incidence of perioperative MI and death [40,41,198,207].

7.2.2 Plexectomy (cardiac sympathetic denervation)

Poor results of aortocoronary bypass grafting were attributed to recurrent vasospasm in the native coronary arteries. Coronary denervation to dilate the distal coronary artery and prevent recurrence of spasm in distal vessel when combined with bypass grafting could improve the results of surgery [6,29,38,208].

Cardiac sympathetic denervation was first described by Arnulf [209]. The modifications of the procedure were described in detail by several authors [41]. It consisted of resection of as much as possible of the major cardiac plexus located anteriorly near the base of the aorta and under the aortic root. Wrisberg's ganglion (sympathetic fibres located in the subaortic and retroaortic areas) should be meticulously dissected and adventitia overlying the aorta peeled off completely.

Some authors reported poor results with this procedure [210]. The most common complication reported with plexectomy was an incomplete denervation [38]. The location and the rich network of fibres explained the difficulty in achieving a satisfactory plexectomy. A meticulous dissection of the rich network of fibres particularly in the posterior plexus would ensure completion and prevent recurrence of symptoms. Complete denervation, in any case, was difficult to achieve through practical operative procedures [41]. Coronary spasm could occur even after total denervation by autotransplantation and the mortality rate for this procedure was very high (25%) and therefore could not be recommended [210].

Recently, unilateral and/or bilateral cervicothoracic sympathetic ganglionectomy for recurrent ventricular tachycardia and prolonged QT interval was reported with good results [211,212].

Plexectomy with CABG offered better results (86%) than CABG alone (61%) and less recurrent attacks (5%) than CABG alone (18%) [198]. A small number of patients refractory to medical management could benefit from plexectomy alone [198].

8 Natural history and prognosis

Prognosis of variant angina was considered to be grave a few decades back [193]. The advent of angiography, better understanding of pathophysiology and availability of calcium channel blockers had a significant impact on the outcome of variant angina. Several factors influenced the long-term course.

Subjects whose initial episodes of vasospastic angina occurred during strenuous efforts were more likely to have a stable but long-term anginal course compared with those experiencing initial episodes of angina at rest [213]. Patients with good initial response to vasodilator therapy were twice more likely to have an event-free clinical course as compared to those with poor initial response to treatment [213].

The degree of coronary artery disease was strong predictor of survival. Survival at 1 year was 99% and at 5 years 94%, for patients with single vessel disease. Survival at 1 and 5 years was only 87% and 77%, respectively, for those with multivessel disease [119,120,214].

In many patients with variant angina, symptoms would disappear spontaneously and the ergonovine test would revert to negative. Treatment with calcium antagonist drugs could probably be safely discontinued in some patients with variant angina; ergonovine test appeared to be helpful in identifying such patients. Longer periods of follow-up were required to confirm that symptoms did not recur [164].

In the absence of significant coronary artery disease, prognosis was good [24,147,215]. Patients with coronary spasm and less than 70% diameter narrowing treated medically had low mortality and morbidity rates. Many patients had long asymptomatic periods and some might be able to stop therapy indefinitely [146,147]. However, Pasternak et al. [39,40] reported that patients without obstructive coronary disease did not necessarily have a benign course. The incidence of cardiac-related deaths was significant in these patients (18%) and a high incidence of life style limiting moderate-to-severe angina [39,40].

Survival for the subpopulation with significant coronary artery disease was significantly better in the Japanese population than in the western populations [216]. However, in the subpopulation without significant coronary artery disease, the prognosis was equally good in western population. If the prevalence of coronary artery disease was corrected for the Japanese population, there would be no difference in the prognosis between the Japanese and the western populations [216].

The prognosis of multivessel spasm was believed to be poor. Anginal attacks due to sequential and simultaneous multivessel spasm were more dangerous than those involving single-vessel spasm or migratory multivessel spasm [217].

9 Conclusion

Patients with variant angina represent a diagnostic and therapeutic dilemma. It was usually suspected from history and substantiated by finding a transient ST segment elevation during a spontaneous attack of angina with good exercise capacity [5]. It should also be considered when rest angina occurred after bypass surgery, particularly if exertional angina was absent and grafts were patent [146].

Provocative tests appeared to be a sensitive method of identifying patients of vasospastic angina but the specificity was less clear. A widespread use of these tests in patients with undiagnosed chest pain syndromes would not be currently recommended.

Long-term management of variant angina depends on the status of the coronary arteries. Patients with isolated coronary artery spasm could have a good clinical outcome with medical management. In cases refractory to medical management, bypass grafting using cardiopulmonary bypass and internal mammary artery conduit was reasonable option [55]. In presence of anatomical narrowing of the coronary arteries, a combined approach involving coronary artery bypass grafting, surgical denervation of the heart and postoperative use of calcium antagonists might give the best results in terms of pain relief and avoidance of recurrence. Cessation of smoking will also reduce the incidence of recurrence.

In survivors of cardiac arrest with no underlying heart disease, ergonovine test should be used to identify patients with coronary spasm and treatment with calcium channel blockers should be initiated. Implantable cardioverter defibrillators, though not needed for most cases, are a good therapeutic option in those with history of life-threatening ventricular arrhythmias.

Variant angina is characterised by considerable spontaneous variability of symptoms making interpretation of results difficult. Spontaneous remission is a frequent outcome of variant angina and should be taken into consideration while evaluating long-term results of therapeutic options [103]. In the absence of data from randomised trials, it is difficult to draw definite conclusions.