- Email: [email protected]
THE MYOCARDIUM IN ANKYLOSING SPONDYLITIS
Saturday 2 May
THE MYOCARDIUM IN ANKYLOSING
SPONDYLITIS A
Clinical, Echocardiographic, and Histopathological Study
DEREK G. GIBSON2 DERRICK A. BREWERTON1 TERENCE J. JONES3 DAVID H. GODDARD1 RICHARD B. MOORE2 COLIN T. PEASE1 PETER A. REVELL3 LEONARD M. SHAPIRO2 KENNETH V. SWETTENHAM3
Department of Rheumatology, Westminster Hospital, London SW1P 2AP;1 Cardiac Department, Brompton Hospital, London SW3 6HP;2 and Department of Morbid Anatomy, London Hospital, E1 1BB3 Cardiac function was investigated in men with ankylosing spondylitis (AS) age 21-65 years who had no cardiorespiratory symptoms or known abnormalities of heart or lungs. Chest radiographs and standard electrocardiograms were normal in 73 of 74 subjects. In echocardiographs of 30 men, left atrial size and left ventricular cavity size and wall thickness were normal. Minor abnormalities in the valve roots were present in 3 older men. Early diastolic abnormalities of the left ventricle were demonstrated in 16 of 30 subjects. This finding was confirmed by repetition of the echocardiography a year later in 15 subjects and by comparison of 11 probands with their healthy brothers. Myocardial tissue obtained at necropsy from 28 AS patients without ischaemic or valvular heart disease or hypertension was studied. A mild, diffuse increase of interstitial connective tissue was seen but there was no
Summary
inflammatory change
amyloid. Computerised image analysis showed 30·7% interstitial reticulin compared with 17·7% in age/sex matched controls (p < 0·0001). or
Introduction IN 1977, Radford et all reported a study of 836 people with ankylosing spondylitis (AS) followed for a mean of 13 years. There were 112 deaths instead of the 63 expected-a significant excess. Although that study did not point particularly to heart disease, it prompted the present
investigation.
1987
Aortitis and aortic regurgitation have been reported in up to 10% of patients with AS.2.3 Minor aortic root abnormalities have been demonstrated on echocardiography." Atrioventricular block has been recorded in 95% of patients in sixteen reports.7-9 Using 24-hour Holter monitoring, Thomsen et a19 found electrocardiographic (ECG) "abnormalities" in 12 of 54 patients with AS. Disorders of the myocardium are customarily said to be uncommon, but the evidence is sparse apart from ECG studies. In previous histopathological reports3,10-12 the findings included aortic inflammation with adventitial scarring, and varying degrees of focal leucocyte infiltration and medial hypertrophy of small arteries usually attributed to the consequences of valvular lesions rather than a primary disorder of the myocardium. By contrast with AS, during the acute phases of the two closely associated diseases Reiter’s disease and yersinia arthritis, myocarditis and pericarditis have been common features in some studies.13,14 In AS, cardiac enlargement and failure was first reported by Graham and Smythe.2 Ribeiro et al15 studied 42 men (28 with AS and 14 with Reiter’s disease) by non-invasive techniques, including echocardiography, and found a dilated, poorly contracting left ventricle with abnormal systolic time intervals in 5 men with AS and 2 with Reiter’s disease who did not have valve disease. These workers postulated an abnormality of the myocardium.
Patients and Methods 74 men who fulfilled the New York Criteria for AS/6 and who had no relevant cardiorespiratory symptoms or disease of heart or lungs, attended for clinical examination, chest radiography, and standard 12-lead resting ECG. From these 74, 27 aged 21-40 years and 29 aged 41-65 years were selected for echocardiography. The younger age-group was confined to those with minimal spinal restriction or deformity to ensure that the recordings were not influenced by spinal disease. Controls (age matched to the nearest year) were men who had no known disease and were not particularly athletic but were attending a fitness clinic. M-mode and crosssectional echocardiograms were performed. Although these echocardiograms were all of a standard suitable for routine clinical purposes, for research it was decided to include only 30 patients
8540
996 with
echocardiograms of high resolution allowing accurate (This proportion accepted was similar to that in
measurements.
other studies and was not influenced by the presence of AS. Those excluded did not differ in age, spinal deformity, or activity of disease.) Of the 30 men remaining for research, 16 were aged 21-40 (mean 32 years) and 14 were 41-65 (mean 52 years). In the younger group 7 had active disease requiring non-steroid anti-inflammatory drugs; and 9 had mild symptoms and were not being treated. None had had radiotherapy. In the older group, 6 men had only mild restriction of spinal movement while 8 had moderate restriction in the thoracolumbar spine. Only 1 had active disease requiring treatment. 3 had had radiotherapy. After the initial study, 15 men with AS attended for repeat echocardiography 12-18 months later. 11 men with AS and their healthy brothers were assessed and had echocardiograms. 3 of the probands had participated in the initial study and 8 had not. The mean ages were 42 years in the probands and 41 years in the healthy brothers. 10 of the original 30 men with AS who had had echocardiography (5 with normal results and 5 with increased MD-MVO) had 24-hour Holter monitoring. The M-mode and cross-sectional echocardiograms were performed with the subjects in the partial left-lateral position. 17 M-mode echocardiograms of the left ventricle (at the level of the mitral valve tips and showing clear continuous echoes from the septum and posterior wall) were recorded at 10 cm per second. Left atrial size was measured (in cm) at end-systole, and posterior wall and septal thickness at end-diastole. The echocardiograms were digitised manually; and from the recordings the following measurements were made and their means calculated from five cardiac cycles: (1) systolic and diastolic left ventricular cavity dimensions in cm; (2) minimum left ventricular dimension to mitral valve opening in ms (MD-MVO); (3) interval from phonocardiographic A2 to mitral valve opening in ms (IVR); (4) peak rate of increase of cavity dimensions during early diastole in cm/s (PRF); and (5) peak rate of posterior wall thinning during early diastole in cm/s (PRT). Forty post-mortem reports of known cases of AS were reviewed. They were obtained from the files of the London Hospital, Hammersmith Hospital, University College Cardiff, and Manchester Royal Infirmary, each of which has a long-standing international reputation for its clinical work on AS and used standard diagnostic criteria throughout the period when the cases were investigated. The four pathology departments follow a similar routine in retaining histological sections of formalin-fixed paraffinembedded myocardium from the left ventricle. Detailed clinical summaries and post-mortem findings were available in all cases, but not the radiographs. 12 cases were excluded on grounds of systemic hypertension before death, ischaemic heart disease (with narrowing of a coronary artery by more than 75 % of its original diameter), or valvular disease. After these exclusions there were 24 male and 4 female cases of AS. 3 men and 1 woman had received radiotherapy for their AS, while 2 further men had had radiotherapy to the thoracic region for carcinoma of the bronchus and thyroid respectively. 16 patients with AS had a probable cardiorespiratory cause of deathrespiratory failure with or without bronchopneumonia (7), congestive cardiac failure (6), and unexpected postoperative death (3). The remaining patients died with carcinomatosis (5), renal failure (3, including 2 with amyloidosis), cerebral infarction (2), and 1 each with haemolytic anaemia and acute leukaemia. The controls were drawn from the post-mortem files of the London Hospital and were sex/age matched (to the nearest 5 years). None had any clinical or pathological evidence of cardiovascular or connective tissue disease. 2 men had received radiotherapy for carcinoma of the bronchus and oesophagus, respectively. 4 patients had received cytotoxic therapy (including doxorubicin) for leukaemia; and doxorubicin had been used in the treatment of a further patient with a hepatoma. Of the controls 10 died of carcinoma, 6 of renal failure, 4 of leukaemia, 2 of hepatic cirrhosis, and 1 each of meningitis, bronchopneumonia, perinephric abscess, hepatoma, astrocytoma, and cerebral aqueduct stenosis. Blocks of tissue were taken either from the posterior wall of the left ventricle beneath the mitral valve (mitral posterior) or from the
Fig 1-MD-MVO and IVR in ms for controls and patients with AS, aged 21-40 and 41-65 years. Shaded
areas
indicate 95% confidence intervals in controls.
lateral free wall of the left ventricle, with the exception of three AS samples and one control which included septal myocardium. Subsequent examination of the results in both AS cases and controls showed that the site in the heart made no difference, there being no systematic bias for a particular site. Without knowledge of the diagnosis, sections from AS cases and controls were examined histologically by use of haematoxylin and eosin, congo red (for amyloid), van Gieson (for connective tissue), and reticulin stains. The reticulin preparations were performed with counterstain, and also without counterstain so that the cell nuclei and cytoplasm were unstained enabling image analysis. The sections stained for reticulin were examined with a Quantimet 720 computerised image analyser, again without knowledge of their origin. Each microscope field was converted to a television image which was then analysed automatically after accurate densitometric adjustment of the contrast between reticulin and all other tissue. The amount of reticulin was measured and expressed as a percentage of the total tissue. Ten fields from each section were examined at a magnification of 250, these conditions having been shown to give reproducible results. The results of the first echocardiographic study and the brother study were analysed with the Wilcoxon rank paired test (twotailed). When 15 patients were studied a second time the two sets of results were compared by means of the Wilcoxon rank paired test and the Spearman rank correlation coefficient (two-tailed). The percentages of reticulin in AS cases and controls were compared by the Mann-Whitney test.
Results In the cardiac
preliminary study, the 74 men with AS had no abnormality on clinical examination, chest radiography, or standard 12-lead resting ECG, apart from 1 man with minor ST-segment changes. On echocardiography, the left atrial size, the maximum and minimum values of left ventricular cavity size, and wall thickness were normal in all men with AS (compared with controls). In 2 older patients there was mitral annular calcification. 1 man with AS aged 45 years had dilatation of the aortic root and mild aortic regurgitation, with minor ST-segment changes. Apart from this man, all patients with
997 ECHOCARDIOGRAPHIC FINDINGS IN MATCHED CONTROLS VERSUS PATIENTS AND HEALTHY BROTHERS VERSUS PATIENTS
are derived from the Wilcoxon rank paired test (two-tailed) and not from the means.
p values
normal-looking aortic valves, normal fractional shortening, and no pronounced abnormalities of regional wall motion. The important abnormalities on echocardiography were AS had
in diastolic ventricular function. In the controls the 95% confidence intervals were as follows: for age 21-40 MD-MVO 0-20 ms, IVR 50-80 ms, PRF 20-11 cm/s, and PRT 12-5 cm/s; for age’ 41-65 MD-MVO 0-25 ms, IVR 45-90 ms, PRF 21-12 cm/s, and PRT 11-4 cm/s. According to these criteria, which conformed with the results in controls in other studies at the Brompton Hospital and National Heart Hospital, the diastolic ventricular function was abnormal in 7 of the 16 younger men with AS and in 9 of the 14 older men (fig 1). In the age group 21-40, the mean MD-MVO was 27 ms compared with 8 ms in the controls (p<002), and the mean IVR was 70 ms compared with 65 ms in the controls (not significant [NS]). 3 men had increases of both MD-MVO and IVR and 4 men had increases of MD-MVO only. The PRF and/or PRT were decreased slightly in 5 men (NS). In the age group 41-65, the mean MD-MVO was 32 ms compared with 10 ms in the controls (p < 0-005) and the mean IVR was 91 ms compared with 68 ms in the controls (p < 002). 6 men had increases of both MD-MVO and IVR, 2 had increases of MD-MVO alone, and 1 had an increase of IVR alone. Of the 3 men who had had radiotherapy, 2 had normal results and 1 had increases of both MD-MVO and IVR. Again, PRF and/or PRT were decreased in 5 men. When all the patients were compared with their controls (table), there were significant increases of MD-MVO (p=00001) and IVR (p=0007), and a decrease in PRT (p < 0-02). There was no correlation
Fig 3-Pereentage cases
reticulin
Fig 2-Low-power photomicrographs to show reticulin content of the myocardium in: (a) control (15% retieulin), (b) AS (30% reticulin). In both photomicrographs
=
mobility, spinal deformity, or treatment. When 15 men with AS were re-examined by echocardiography after 12-18 months, the means for MD-MVO, IVR, PRF, and PRT were virtually unaltered. The Wilcoxon rank paired test was not significantly altered and the Spearman rank correlation coefficient (two-tailed) In was p < 0-00002 for each set of measurements. MD-MVO and IVR only one measurement altered more than 5 ms: in a man aged 44 with AS the IVR increased from 75 (normal) to 100 ms. When 11healthy brothers were compared with probands who had AS, the pattern of results was the same as in the initial echocardiographic study. In the probands the MD-MVO and/or IVR was prolonged in 5 of 6 aged 21-40 and in 4 of 5 aged 41-65. The means for MD-MVO and IVR were 9 and 70 ms in the healthy brothers compared with 33 and 89 ms in the probands (p < 0 005 and < 0006). The means for PRF and PRT were
of myocardium in individual males, 0 females) against age in =
years.
Linear
regressions
are
120.
between any of these abnormalities of diastolic ventricular function and duration of disease, activity of disease, spinal
content
of AS and controls (0
magnification is
of reticulin percentage against age.
998 17 and 8 cm/s in the brothers compared with 14 and 6 cm/s in the probands (p < 0-01 and NS). 5 of the healthy brothers had HLA-B27 and 6 did not: all had normal echocardiograms. The 24-hour Holter monitoring was normal in the 10 patients investigated. In the study of necropsy material the mean age at death was 54 4 years for AS and 54 5 years for controls. In cases of AS, conventional microscopy of the myocardium showed no inflammatory changes. Stains for amyloid were uniformly negative. There was a mild diffuse increase of interstitial connective tissue between individual muscle cells and around muscle bundles, with a corresponding increase in reticulin fibres (fig 2). Computerised image-analysis showed a significantly greater amount of interstitial reticulin in AS cases than in controls (fig 3): AS mean 30-7% (SD 7-3%), controls 17-7% (SD 6-3%); Mann-Whitney test p < 0-0001. The results in the 4 women were similar to those in men, both in AS cases and in controls. The linear regressions in fig 3 are of reticulin percentage against age. The reticulin increased with age in both AS cases and controls and the rates of increase with age did not differ between these groups. Among AS cases the amount of reticulin was not greater in those who had died of a cardiac or respiratory cause, who had died postoperatively, or who had had radiotherapy; and in controls the reticulin content was not
influenced by
radiotherapy or cytotoxic therapy. Discussion
In the clinical investigation the initial selection of patients
critical; and the results may differ from those of some other workers because of our decisions to exclude all patients with clinical evidence of heart or lung disease and to ensure the inclusion of a well-defined group of men aged 21-40 with good spinal movements and little deformity. The 3 patients with minor valve root abnormalities were in the older age group; and only 1 patient (with previously unrecognised aortic regurgitation) had minor ECG changes. The other abnormalities of the aortic root reported by Tucker et al4 were not seen. In all our patients investigated by echocardiography, the left atrial size, maximum and minimum left ventricular cavity size, and wall thickness were normal: we did not encounter patients with dilated poorly contracting ventricles as described by Ribeiro et al15 The abnormality we identified was different and has not previously, been reported: it represents a functional disturbance of the left ventricle. The abnormalities we observed on echocardiography were predominantly in early diastole (MD-MVO and IVR). They are not indicative of coronary artery disease, conduction defects, valve disease, or amyloidosis, nor do they resemble the usual findings in myocarditis or systemic sclerosis. Increases of MD-MVO and IVR may be seen in patients with left ventricular hypertrophy secondary to hypertension or hypertrophic obstructive . cardiomyopathy, 18 but this is not the explanation in our subjects because they did not have ventricular hypertrophy as judged by echocardiography, ECG, or chest radiography. Similar disturbances of left ventricular function without ventricular hypertrophy have been observed in men with diabetes mellitus aged 16-60 years, particularly in those with microangiopathy and severe disease.19It was the conclusion that the most likely explanation of our observations in AS was excess myocardial connective tissue that led to reappraisal of the histology of the heart in AS post mortem. The histopathological study showed that, in patients with AS over the age of 30 years, there is commonly a diffuse was
increase of myocardial interstitial connective tissue, even in the absence of clinical or pathological evidence of other cardiac abnormalities. As far as can be assessed from post-mortem material, the rate of increase of myocardial connective tissue with age is no greater in AS than in controls. There was no relation between cardiorespiratory or sudden death and the amount of myocardial connective tissue; and radiotherapy was not a contributory factor in either AS patients or controls. Ankylosing spondylitis is a common disease. It is of considerable concern that many apparently fit young men with mild spinal disease have abnormal diastolic function, but at present there is no evidence that this abnormality has clinical consequences or indicates a bad prognosis. There is now a strong case for a long-term study of a large number of men and women aged 35--45 who have undergone comprehensive investigation of the heart by non-invasive techniques. This seems to be the only means of determining which abnormalities indicate subsequent increased morbidity or mortality and whether prophylactic treatment is feasible. There is also need for a multicentre, prospective clinical/pathological study that would include the examination of snap-frozen sections of the myocardium with modem histochemical and immunohistochemical methods.
follow-up
We
are
grateful
to
the
rheumatologists
and
pathologists
at the London
Hospital, Hammersmith Hospital, University College Hospital Cardiff, and Manchester Royal Infirmary for their enthusiastic cooperation, and to Dr David Preston and Miss Rosalind Wood, Medical Computer Centre, Westminster Hospital, for their statistical analysis of the results. Dr R. Sutton kindly supervised the 24-hour Holter monitoring.
Correspondence should be addressed to D. A. B. REFERENCES 1. Radford EP, Doll
R, Smith PE. Mortality among patients with ankylosing spondylitis given Xray therapy. N Engl J Med 1977; 297: 572-76. Graham DC, Smythe HA The carditis and aortitis of ankylosing spondylitis Bull not
2.
Rheum Dis 1958; 9: 171-74. 3. Kinsella TD, Johnson LG, Sutherland RI. Cardiovascular manifestations of ankylosing spondylitis. Can Med Assoc J 1974; 111: 1309-11. 4 Tucker CR, Fowles RE, Calin A, Popp RL. Aortitis in ankylosing spondylitis Early detection of aortic root abnormalities with two dimensional echocardiography. Am J Cardiol 1982; 49: 680-86. 5. Thomas D, Hill W, Geddes R, et al. Early detection of aortic dilatation in ankylosing spondylitis using echocardiography. Aust NZ J Med 1982; 12: 10-13. 6. Labresh KA, Lally EV, Sharma SC, Ho G. Two-dimensional echocardiographic detection of preclinical aortic root abnormalities in rheumatoid variant diseases. Am J Med 1985; 78: 908-12. 7. Takkunen J, Vuopala U, Isomaki H. Cardiomyopathy in ankylosing spondylitis. Ann Clin Research 1970; 2: 106-12. 8. Bergfeldt L, Edhag O, Vallin H. Cardiac conduction disturbances, an underestimated manifestation in ankylosing spondylitis. Acta Med Scand 1982; 212: 217-23. 9. Thomsen NH, Hørslev-Petersen K, Beyer JM. Ambulatory 24-hour continuous electrocardiographic monitoring in 54 patients with ankylosing spondylitis. Am Heart J 1986; 7: 240-46. 10. Bergfeldt L, Edhag O, Rajs J. HLA-B27 associated heart disease. Clinicopathologic study of three cases. Am J Med 1984; 77: 961-67. 11. Bulkley BH, Roberts WC. Ankylosing spondylitis and aortic regurgitation. Description of the characteristic cardiovascular lesion from study of eight necropsy patients. Circulation 1973; 48: 1014-27. 12. Davidson P, Baggenstoss AH, Slocumb CH, Daugherty GW. Cardiac and aortic lesions in rheumatoid spondylitis. Proc Staff Meet Mayo Clin 1963; 38: 427-35. 13. Paronen I. Reiter’s disease. A study of 344 cases observed in Finland. Acta Med Scand 1948; 131 (suppl 212): 1-114. 14. Laitinen O, Leirisalo M, Allander E. Rheumatic fever and yersinia arthritis. Criteria and diagnostic problems in a changing disease pattern. Scand J Rheumatol 1975; 4: 145-57. 15. Ribeiro P, Morley KD, Shapiro LM, Garnett RAF, Hughes GRV, Goodwin JF. Left ventricular function in patients with ankylosing spondylitis. Acta Med Scand 1982; 212: 217-23. 16. Bennett PH, Wood PHN. Population studies of the rheumatic diseases Proceedings of the Third International Symposium. Amsterdam. Amsterdam: Excerpta Medica, 1968: 456-57. 17. Gibson DG, Brown DJ. Measurement of instantaneous left ventricular dimension and filling rate in man using echocardiography. Br Med J 1973; 35: 1141-49. 18. Shapiro LM, Moore RB, Logan-Sinclair BB, Gibson DG. Relation of regional echo amplitude to left ventricular function and electrocardiogram in left ventricular 19.
hypertrophy. Br Heart J 1984; 52: 99-105. Shapiro LM. Echocardiographic features of impaired ventricular function in diabetes mellitus. Br Heart J 1982; 47: 439-44.
THE MYOCARDIUM IN ANKYLOSING
SPONDYLITIS A
Clinical, Echocardiographic, and Histopathological Study
DEREK G. GIBSON2 DERRICK A. BREWERTON1 TERENCE J. JONES3 DAVID H. GODDARD1 RICHARD B. MOORE2 COLIN T. PEASE1 PETER A. REVELL3 LEONARD M. SHAPIRO2 KENNETH V. SWETTENHAM3
Department of Rheumatology, Westminster Hospital, London SW1P 2AP;1 Cardiac Department, Brompton Hospital, London SW3 6HP;2 and Department of Morbid Anatomy, London Hospital, E1 1BB3 Cardiac function was investigated in men with ankylosing spondylitis (AS) age 21-65 years who had no cardiorespiratory symptoms or known abnormalities of heart or lungs. Chest radiographs and standard electrocardiograms were normal in 73 of 74 subjects. In echocardiographs of 30 men, left atrial size and left ventricular cavity size and wall thickness were normal. Minor abnormalities in the valve roots were present in 3 older men. Early diastolic abnormalities of the left ventricle were demonstrated in 16 of 30 subjects. This finding was confirmed by repetition of the echocardiography a year later in 15 subjects and by comparison of 11 probands with their healthy brothers. Myocardial tissue obtained at necropsy from 28 AS patients without ischaemic or valvular heart disease or hypertension was studied. A mild, diffuse increase of interstitial connective tissue was seen but there was no
Summary
inflammatory change
amyloid. Computerised image analysis showed 30·7% interstitial reticulin compared with 17·7% in age/sex matched controls (p < 0·0001). or
Introduction IN 1977, Radford et all reported a study of 836 people with ankylosing spondylitis (AS) followed for a mean of 13 years. There were 112 deaths instead of the 63 expected-a significant excess. Although that study did not point particularly to heart disease, it prompted the present
investigation.
1987
Aortitis and aortic regurgitation have been reported in up to 10% of patients with AS.2.3 Minor aortic root abnormalities have been demonstrated on echocardiography." Atrioventricular block has been recorded in 95% of patients in sixteen reports.7-9 Using 24-hour Holter monitoring, Thomsen et a19 found electrocardiographic (ECG) "abnormalities" in 12 of 54 patients with AS. Disorders of the myocardium are customarily said to be uncommon, but the evidence is sparse apart from ECG studies. In previous histopathological reports3,10-12 the findings included aortic inflammation with adventitial scarring, and varying degrees of focal leucocyte infiltration and medial hypertrophy of small arteries usually attributed to the consequences of valvular lesions rather than a primary disorder of the myocardium. By contrast with AS, during the acute phases of the two closely associated diseases Reiter’s disease and yersinia arthritis, myocarditis and pericarditis have been common features in some studies.13,14 In AS, cardiac enlargement and failure was first reported by Graham and Smythe.2 Ribeiro et al15 studied 42 men (28 with AS and 14 with Reiter’s disease) by non-invasive techniques, including echocardiography, and found a dilated, poorly contracting left ventricle with abnormal systolic time intervals in 5 men with AS and 2 with Reiter’s disease who did not have valve disease. These workers postulated an abnormality of the myocardium.
Patients and Methods 74 men who fulfilled the New York Criteria for AS/6 and who had no relevant cardiorespiratory symptoms or disease of heart or lungs, attended for clinical examination, chest radiography, and standard 12-lead resting ECG. From these 74, 27 aged 21-40 years and 29 aged 41-65 years were selected for echocardiography. The younger age-group was confined to those with minimal spinal restriction or deformity to ensure that the recordings were not influenced by spinal disease. Controls (age matched to the nearest year) were men who had no known disease and were not particularly athletic but were attending a fitness clinic. M-mode and crosssectional echocardiograms were performed. Although these echocardiograms were all of a standard suitable for routine clinical purposes, for research it was decided to include only 30 patients
8540
996 with
echocardiograms of high resolution allowing accurate (This proportion accepted was similar to that in
measurements.
other studies and was not influenced by the presence of AS. Those excluded did not differ in age, spinal deformity, or activity of disease.) Of the 30 men remaining for research, 16 were aged 21-40 (mean 32 years) and 14 were 41-65 (mean 52 years). In the younger group 7 had active disease requiring non-steroid anti-inflammatory drugs; and 9 had mild symptoms and were not being treated. None had had radiotherapy. In the older group, 6 men had only mild restriction of spinal movement while 8 had moderate restriction in the thoracolumbar spine. Only 1 had active disease requiring treatment. 3 had had radiotherapy. After the initial study, 15 men with AS attended for repeat echocardiography 12-18 months later. 11 men with AS and their healthy brothers were assessed and had echocardiograms. 3 of the probands had participated in the initial study and 8 had not. The mean ages were 42 years in the probands and 41 years in the healthy brothers. 10 of the original 30 men with AS who had had echocardiography (5 with normal results and 5 with increased MD-MVO) had 24-hour Holter monitoring. The M-mode and cross-sectional echocardiograms were performed with the subjects in the partial left-lateral position. 17 M-mode echocardiograms of the left ventricle (at the level of the mitral valve tips and showing clear continuous echoes from the septum and posterior wall) were recorded at 10 cm per second. Left atrial size was measured (in cm) at end-systole, and posterior wall and septal thickness at end-diastole. The echocardiograms were digitised manually; and from the recordings the following measurements were made and their means calculated from five cardiac cycles: (1) systolic and diastolic left ventricular cavity dimensions in cm; (2) minimum left ventricular dimension to mitral valve opening in ms (MD-MVO); (3) interval from phonocardiographic A2 to mitral valve opening in ms (IVR); (4) peak rate of increase of cavity dimensions during early diastole in cm/s (PRF); and (5) peak rate of posterior wall thinning during early diastole in cm/s (PRT). Forty post-mortem reports of known cases of AS were reviewed. They were obtained from the files of the London Hospital, Hammersmith Hospital, University College Cardiff, and Manchester Royal Infirmary, each of which has a long-standing international reputation for its clinical work on AS and used standard diagnostic criteria throughout the period when the cases were investigated. The four pathology departments follow a similar routine in retaining histological sections of formalin-fixed paraffinembedded myocardium from the left ventricle. Detailed clinical summaries and post-mortem findings were available in all cases, but not the radiographs. 12 cases were excluded on grounds of systemic hypertension before death, ischaemic heart disease (with narrowing of a coronary artery by more than 75 % of its original diameter), or valvular disease. After these exclusions there were 24 male and 4 female cases of AS. 3 men and 1 woman had received radiotherapy for their AS, while 2 further men had had radiotherapy to the thoracic region for carcinoma of the bronchus and thyroid respectively. 16 patients with AS had a probable cardiorespiratory cause of deathrespiratory failure with or without bronchopneumonia (7), congestive cardiac failure (6), and unexpected postoperative death (3). The remaining patients died with carcinomatosis (5), renal failure (3, including 2 with amyloidosis), cerebral infarction (2), and 1 each with haemolytic anaemia and acute leukaemia. The controls were drawn from the post-mortem files of the London Hospital and were sex/age matched (to the nearest 5 years). None had any clinical or pathological evidence of cardiovascular or connective tissue disease. 2 men had received radiotherapy for carcinoma of the bronchus and oesophagus, respectively. 4 patients had received cytotoxic therapy (including doxorubicin) for leukaemia; and doxorubicin had been used in the treatment of a further patient with a hepatoma. Of the controls 10 died of carcinoma, 6 of renal failure, 4 of leukaemia, 2 of hepatic cirrhosis, and 1 each of meningitis, bronchopneumonia, perinephric abscess, hepatoma, astrocytoma, and cerebral aqueduct stenosis. Blocks of tissue were taken either from the posterior wall of the left ventricle beneath the mitral valve (mitral posterior) or from the
Fig 1-MD-MVO and IVR in ms for controls and patients with AS, aged 21-40 and 41-65 years. Shaded
areas
indicate 95% confidence intervals in controls.
lateral free wall of the left ventricle, with the exception of three AS samples and one control which included septal myocardium. Subsequent examination of the results in both AS cases and controls showed that the site in the heart made no difference, there being no systematic bias for a particular site. Without knowledge of the diagnosis, sections from AS cases and controls were examined histologically by use of haematoxylin and eosin, congo red (for amyloid), van Gieson (for connective tissue), and reticulin stains. The reticulin preparations were performed with counterstain, and also without counterstain so that the cell nuclei and cytoplasm were unstained enabling image analysis. The sections stained for reticulin were examined with a Quantimet 720 computerised image analyser, again without knowledge of their origin. Each microscope field was converted to a television image which was then analysed automatically after accurate densitometric adjustment of the contrast between reticulin and all other tissue. The amount of reticulin was measured and expressed as a percentage of the total tissue. Ten fields from each section were examined at a magnification of 250, these conditions having been shown to give reproducible results. The results of the first echocardiographic study and the brother study were analysed with the Wilcoxon rank paired test (twotailed). When 15 patients were studied a second time the two sets of results were compared by means of the Wilcoxon rank paired test and the Spearman rank correlation coefficient (two-tailed). The percentages of reticulin in AS cases and controls were compared by the Mann-Whitney test.
Results In the cardiac
preliminary study, the 74 men with AS had no abnormality on clinical examination, chest radiography, or standard 12-lead resting ECG, apart from 1 man with minor ST-segment changes. On echocardiography, the left atrial size, the maximum and minimum values of left ventricular cavity size, and wall thickness were normal in all men with AS (compared with controls). In 2 older patients there was mitral annular calcification. 1 man with AS aged 45 years had dilatation of the aortic root and mild aortic regurgitation, with minor ST-segment changes. Apart from this man, all patients with
997 ECHOCARDIOGRAPHIC FINDINGS IN MATCHED CONTROLS VERSUS PATIENTS AND HEALTHY BROTHERS VERSUS PATIENTS
are derived from the Wilcoxon rank paired test (two-tailed) and not from the means.
p values
normal-looking aortic valves, normal fractional shortening, and no pronounced abnormalities of regional wall motion. The important abnormalities on echocardiography were AS had
in diastolic ventricular function. In the controls the 95% confidence intervals were as follows: for age 21-40 MD-MVO 0-20 ms, IVR 50-80 ms, PRF 20-11 cm/s, and PRT 12-5 cm/s; for age’ 41-65 MD-MVO 0-25 ms, IVR 45-90 ms, PRF 21-12 cm/s, and PRT 11-4 cm/s. According to these criteria, which conformed with the results in controls in other studies at the Brompton Hospital and National Heart Hospital, the diastolic ventricular function was abnormal in 7 of the 16 younger men with AS and in 9 of the 14 older men (fig 1). In the age group 21-40, the mean MD-MVO was 27 ms compared with 8 ms in the controls (p<002), and the mean IVR was 70 ms compared with 65 ms in the controls (not significant [NS]). 3 men had increases of both MD-MVO and IVR and 4 men had increases of MD-MVO only. The PRF and/or PRT were decreased slightly in 5 men (NS). In the age group 41-65, the mean MD-MVO was 32 ms compared with 10 ms in the controls (p < 0-005) and the mean IVR was 91 ms compared with 68 ms in the controls (p < 002). 6 men had increases of both MD-MVO and IVR, 2 had increases of MD-MVO alone, and 1 had an increase of IVR alone. Of the 3 men who had had radiotherapy, 2 had normal results and 1 had increases of both MD-MVO and IVR. Again, PRF and/or PRT were decreased in 5 men. When all the patients were compared with their controls (table), there were significant increases of MD-MVO (p=00001) and IVR (p=0007), and a decrease in PRT (p < 0-02). There was no correlation
Fig 3-Pereentage cases
reticulin
Fig 2-Low-power photomicrographs to show reticulin content of the myocardium in: (a) control (15% retieulin), (b) AS (30% reticulin). In both photomicrographs
=
mobility, spinal deformity, or treatment. When 15 men with AS were re-examined by echocardiography after 12-18 months, the means for MD-MVO, IVR, PRF, and PRT were virtually unaltered. The Wilcoxon rank paired test was not significantly altered and the Spearman rank correlation coefficient (two-tailed) In was p < 0-00002 for each set of measurements. MD-MVO and IVR only one measurement altered more than 5 ms: in a man aged 44 with AS the IVR increased from 75 (normal) to 100 ms. When 11healthy brothers were compared with probands who had AS, the pattern of results was the same as in the initial echocardiographic study. In the probands the MD-MVO and/or IVR was prolonged in 5 of 6 aged 21-40 and in 4 of 5 aged 41-65. The means for MD-MVO and IVR were 9 and 70 ms in the healthy brothers compared with 33 and 89 ms in the probands (p < 0 005 and < 0006). The means for PRF and PRT were
of myocardium in individual males, 0 females) against age in =
years.
Linear
regressions
are
120.
between any of these abnormalities of diastolic ventricular function and duration of disease, activity of disease, spinal
content
of AS and controls (0
magnification is
of reticulin percentage against age.
998 17 and 8 cm/s in the brothers compared with 14 and 6 cm/s in the probands (p < 0-01 and NS). 5 of the healthy brothers had HLA-B27 and 6 did not: all had normal echocardiograms. The 24-hour Holter monitoring was normal in the 10 patients investigated. In the study of necropsy material the mean age at death was 54 4 years for AS and 54 5 years for controls. In cases of AS, conventional microscopy of the myocardium showed no inflammatory changes. Stains for amyloid were uniformly negative. There was a mild diffuse increase of interstitial connective tissue between individual muscle cells and around muscle bundles, with a corresponding increase in reticulin fibres (fig 2). Computerised image-analysis showed a significantly greater amount of interstitial reticulin in AS cases than in controls (fig 3): AS mean 30-7% (SD 7-3%), controls 17-7% (SD 6-3%); Mann-Whitney test p < 0-0001. The results in the 4 women were similar to those in men, both in AS cases and in controls. The linear regressions in fig 3 are of reticulin percentage against age. The reticulin increased with age in both AS cases and controls and the rates of increase with age did not differ between these groups. Among AS cases the amount of reticulin was not greater in those who had died of a cardiac or respiratory cause, who had died postoperatively, or who had had radiotherapy; and in controls the reticulin content was not
influenced by
radiotherapy or cytotoxic therapy. Discussion
In the clinical investigation the initial selection of patients
critical; and the results may differ from those of some other workers because of our decisions to exclude all patients with clinical evidence of heart or lung disease and to ensure the inclusion of a well-defined group of men aged 21-40 with good spinal movements and little deformity. The 3 patients with minor valve root abnormalities were in the older age group; and only 1 patient (with previously unrecognised aortic regurgitation) had minor ECG changes. The other abnormalities of the aortic root reported by Tucker et al4 were not seen. In all our patients investigated by echocardiography, the left atrial size, maximum and minimum left ventricular cavity size, and wall thickness were normal: we did not encounter patients with dilated poorly contracting ventricles as described by Ribeiro et al15 The abnormality we identified was different and has not previously, been reported: it represents a functional disturbance of the left ventricle. The abnormalities we observed on echocardiography were predominantly in early diastole (MD-MVO and IVR). They are not indicative of coronary artery disease, conduction defects, valve disease, or amyloidosis, nor do they resemble the usual findings in myocarditis or systemic sclerosis. Increases of MD-MVO and IVR may be seen in patients with left ventricular hypertrophy secondary to hypertension or hypertrophic obstructive . cardiomyopathy, 18 but this is not the explanation in our subjects because they did not have ventricular hypertrophy as judged by echocardiography, ECG, or chest radiography. Similar disturbances of left ventricular function without ventricular hypertrophy have been observed in men with diabetes mellitus aged 16-60 years, particularly in those with microangiopathy and severe disease.19It was the conclusion that the most likely explanation of our observations in AS was excess myocardial connective tissue that led to reappraisal of the histology of the heart in AS post mortem. The histopathological study showed that, in patients with AS over the age of 30 years, there is commonly a diffuse was
increase of myocardial interstitial connective tissue, even in the absence of clinical or pathological evidence of other cardiac abnormalities. As far as can be assessed from post-mortem material, the rate of increase of myocardial connective tissue with age is no greater in AS than in controls. There was no relation between cardiorespiratory or sudden death and the amount of myocardial connective tissue; and radiotherapy was not a contributory factor in either AS patients or controls. Ankylosing spondylitis is a common disease. It is of considerable concern that many apparently fit young men with mild spinal disease have abnormal diastolic function, but at present there is no evidence that this abnormality has clinical consequences or indicates a bad prognosis. There is now a strong case for a long-term study of a large number of men and women aged 35--45 who have undergone comprehensive investigation of the heart by non-invasive techniques. This seems to be the only means of determining which abnormalities indicate subsequent increased morbidity or mortality and whether prophylactic treatment is feasible. There is also need for a multicentre, prospective clinical/pathological study that would include the examination of snap-frozen sections of the myocardium with modem histochemical and immunohistochemical methods.
follow-up
We
are
grateful
to
the
rheumatologists
and
pathologists
at the London
Hospital, Hammersmith Hospital, University College Hospital Cardiff, and Manchester Royal Infirmary for their enthusiastic cooperation, and to Dr David Preston and Miss Rosalind Wood, Medical Computer Centre, Westminster Hospital, for their statistical analysis of the results. Dr R. Sutton kindly supervised the 24-hour Holter monitoring.
Correspondence should be addressed to D. A. B. REFERENCES 1. Radford EP, Doll
R, Smith PE. Mortality among patients with ankylosing spondylitis given Xray therapy. N Engl J Med 1977; 297: 572-76. Graham DC, Smythe HA The carditis and aortitis of ankylosing spondylitis Bull not
2.
Rheum Dis 1958; 9: 171-74. 3. Kinsella TD, Johnson LG, Sutherland RI. Cardiovascular manifestations of ankylosing spondylitis. Can Med Assoc J 1974; 111: 1309-11. 4 Tucker CR, Fowles RE, Calin A, Popp RL. Aortitis in ankylosing spondylitis Early detection of aortic root abnormalities with two dimensional echocardiography. Am J Cardiol 1982; 49: 680-86. 5. Thomas D, Hill W, Geddes R, et al. Early detection of aortic dilatation in ankylosing spondylitis using echocardiography. Aust NZ J Med 1982; 12: 10-13. 6. Labresh KA, Lally EV, Sharma SC, Ho G. Two-dimensional echocardiographic detection of preclinical aortic root abnormalities in rheumatoid variant diseases. Am J Med 1985; 78: 908-12. 7. Takkunen J, Vuopala U, Isomaki H. Cardiomyopathy in ankylosing spondylitis. Ann Clin Research 1970; 2: 106-12. 8. Bergfeldt L, Edhag O, Vallin H. Cardiac conduction disturbances, an underestimated manifestation in ankylosing spondylitis. Acta Med Scand 1982; 212: 217-23. 9. Thomsen NH, Hørslev-Petersen K, Beyer JM. Ambulatory 24-hour continuous electrocardiographic monitoring in 54 patients with ankylosing spondylitis. Am Heart J 1986; 7: 240-46. 10. Bergfeldt L, Edhag O, Rajs J. HLA-B27 associated heart disease. Clinicopathologic study of three cases. Am J Med 1984; 77: 961-67. 11. Bulkley BH, Roberts WC. Ankylosing spondylitis and aortic regurgitation. Description of the characteristic cardiovascular lesion from study of eight necropsy patients. Circulation 1973; 48: 1014-27. 12. Davidson P, Baggenstoss AH, Slocumb CH, Daugherty GW. Cardiac and aortic lesions in rheumatoid spondylitis. Proc Staff Meet Mayo Clin 1963; 38: 427-35. 13. Paronen I. Reiter’s disease. A study of 344 cases observed in Finland. Acta Med Scand 1948; 131 (suppl 212): 1-114. 14. Laitinen O, Leirisalo M, Allander E. Rheumatic fever and yersinia arthritis. Criteria and diagnostic problems in a changing disease pattern. Scand J Rheumatol 1975; 4: 145-57. 15. Ribeiro P, Morley KD, Shapiro LM, Garnett RAF, Hughes GRV, Goodwin JF. Left ventricular function in patients with ankylosing spondylitis. Acta Med Scand 1982; 212: 217-23. 16. Bennett PH, Wood PHN. Population studies of the rheumatic diseases Proceedings of the Third International Symposium. Amsterdam. Amsterdam: Excerpta Medica, 1968: 456-57. 17. Gibson DG, Brown DJ. Measurement of instantaneous left ventricular dimension and filling rate in man using echocardiography. Br Med J 1973; 35: 1141-49. 18. Shapiro LM, Moore RB, Logan-Sinclair BB, Gibson DG. Relation of regional echo amplitude to left ventricular function and electrocardiogram in left ventricular 19.
hypertrophy. Br Heart J 1984; 52: 99-105. Shapiro LM. Echocardiographic features of impaired ventricular function in diabetes mellitus. Br Heart J 1982; 47: 439-44.