ROUTINE CORONARY ANGIOGRAPHY FOR EFFORT ANGINA

837 bismuth subcitrate (CBS) (19/22, 86%). We agree that our patients are unique, as Lam says, since all patients are unique. Every effort was taken to keep the trial blind. Follow-up was timed 2 weeks after therapy ceased (so that mouth-staining would not be seen by the endoscopist) and communication was avoided between the clinician and the patient before gastroscopy. More importantly the pathologist and microbiologist were independent; they had no knowledge of the therapy and neither the clinician nor the patient knew of the presence or absence of bacteria. Thus we do not agree that the title is misleading. Lam mentions that our definition of ulcer relapse includes "any recurrence of ulcer symptoms". This is almost correct--our patients were treated as in normal practice. If they complained of symptoms requiring therapy, they were considered to have relapsed and usually (32/43, 74%) an ulcer was found on gastroscopy. We thought it unreasonable to call such patients "successfully treated" if they had a recurrence of symptoms but did not have an ulcer crater at

endoscopy.

We agree that "bismuth not cimetidine is associated with clearance of campylobacter", but if Lam really believes we did not "safeguard against investigator or patient bias", we suggest he considers the following. There is no way to demonstate CBS therapy on histology or microbiology, unless Lam thinks that the absence of gastritis and campylobacter, only seen with CBS therapy, causes observer bias. All bias can thus be eliminated by considering only the group of patients treated with CBS. This shows 22 CP+ve patients, 10 (45%) healed at follow-up and 3 (14%) still healed at 12 months; compare this with 24 CP-ve patients, 22 (92%) healed and 17 (71 %) still healed at 12 months. Tinidazole was not intended to "differentiate cimetidine and bismuth", nor for that matter was our study. If Lam is interested in tinidazole, he should examine our results for CBS with or without tinidazole, which show an improvement with tinidazole. We were investigating the effect of the eradication of C pylori, not the effect of cimetidine, CBS, or tinidazole per se--our results should be viewed from this aspect. The distribution of campylobacter is patchy, as stated by Lam, but only at a microscopic level and, with improved methods, we rarely have contradictory results. In this series no CP + ve patient gave negative histological findings and culture without treatment and only 1 patient gave two successive negative biopsy specimens followed by a Cpylori positive culture (probably a true re-infection). The series includes about 450 biopsies on 100 patients. Under these circumstances we consider "eradication" a correct and justified

and the T-cell receptor gene rearrangements.1-3 The lesions tend to affect extranodal sites such as the upper respiratory tract, lung, and skin, and synchronous or metachronous involvement of more than one of these sites may occursThey differ histologically from mycosis fungoides in the prominent vascular invasion, frequent necrosis, more deeply located dermal infiltrate, infrequent epidermal invasion, and rarity of cerebriform cells.3 Your editorial equates midline granuloma with polymorphic reticulosis but "midline granuloma" refers to a progressive, necrotising lesion affecting the nose and adjacent structures, and a wide variety of diseases can produce that clinical picture, including infection, vasculitis, and angiocentric lymphoproliferative lesions.2,4 On the other hand, "polymorphic reticulosis" is a purely histological designation. AILD is an altogether different entity clinically and histologically,’’ and morphologically and cytologically it is distinct from either lymphomatoid granulomatosis and mycosis fungoides. Many cases reported as AILD may really be peripheral T-cell lymphoma (immunoblastic lymphadenopathy-like T-cell

immunophenotype,

lymphoma).s Although Scott and Robb-Smith did not label HMR as malignant in their original description, Robb-Smith subsequently considered it a proliferation of malignant and actively phagocytic histiocytes.6 Immunohistochemical and genotypic studies have shown that most malignant "histiocytic" proliferations are T or B cell malignancies.7 Malignant true histiocytic diseases do exist but most cases satisfying the diagnostic criteria of HMR are probably large-cell lymphomas, such as anaplastic Ki-1or peripheral T-cell lymphoma, associated with reactive haemophagocytic syndrome.8,9 Angiocentric lymphoproliferative lesions, as in the case reported by Whittaker and colleagues are also not uncommonly complicated by the haemophagocytic syndrome.1.3 What has been reported as HMR has probably included reactive haemophagocytic syndrome as well, especially in reports before the description of this benign syndrome. Institute of Pathology, Queen Elizabeth Hospital, Kowloon, Hong Kong Institute of Pathology, Caritas Medical Centre, Hong Kong

1.

C. S. NG

Lipford EH, Margolick JB, Longo DL, et al. Angiocentric immunoproliferative lesions. a clinicopathologic spectrum of post-thymic T-cell proliferations. Blood

1988; 72: 1674-81. 2. Chan JKC, Ng CS, Lau WH, et al. Most nasal/nasopharyngeal

term.

Our work was designed to show the effect of the eradication of C pylori on duodenal ulcer relapse. It was never intended as a drug trial. Cimetidine is an ulcer-healing agent but will not eradicate C pylori. Our results suggest that eradication of C pylori is associated with a dramatic reduction in the relapse of healed ulcers.

J. K. C. CHAN

3.

lymphomas are T-cell neoplasms Am J Surg Pathol 1987; 11: 418-29. Chan JKC, Ng CS, Ngan KC, et al. Angiocentric T-cell lymphoma of the skin, an aggressive lymphoma distinct from mycosis fungoides. Am J Surg Pathol 1988; 12:

861-76. 4. Costa J, Delacretaz F. The midline granuloma syndrome. Pathol Annu 1986; 21: 159-71 5. Nathwani BN, Brynes RK. Reactive unmunoblastic proliferations. Semin Diag Pathol

1988; 5: 317-28. 6. Robb-Smith AHT, Taylor CR Lymph node biopsy. London: Miller Heyden, 1981 137-39. 7. Weiss LM, Trela MJ, Cleary ML. Frequent immunoglobulin and T-cell receptor rearrangement in ’histiocytic’ neoplasms. Am J Pathol 1985; 121: 369-73. 8. Chan JKC, Ng CS, Hui PK, et al. Anaplastic large cell Ki-1 lymphoma, delineation of :

Internal Medicine, Health Sciences Center, University of Virginia,

J. MARSHALL

Charlottesville, Virginia 22908, USA

BARRY

Royal Perth Hospital, Perth, Western Australia

J. ROBIN WARREN C. STEWART GOODWIN

TRENDS IN LYMPHOMA DIAGNOSIS

SiR,—Your Feb 4 editorial (p 249) erroneously suggests lymphomatoid granulomatosis, midline granuloma, angioimmunoblastic lymphadenopathy AILD), histiocytic medullary reticulosis (HMR), and mycosis fungoides are related, and that the skin lesions seen in these conditions are similar. In fact, all these conditions are very different clinically and pathologically, except for lymphomatoid granulomatosis and polymorphic reticulosis (midline granuloma). Lymphomatoid granulomatosis and polymorphic reticulosis are similar if not identical.’ They are angiocentric and angiodestructive that

lymphoproliferative lesions which may evolve into a more monomorphic picture of angiocentric lymphoma. Despite their polymorphous cellular composition, they are held to be neoplastic diseases by virtue of their ability to metastasise, aberrant T-cell

two morphological types. Histopathology (in press). 9 Chan JKC, Ng CS, Law CK, et al. Reactive haemophagocytic syndrome, a study of 7 fatal cases. Pathology 1987; 19: 43-50. 10. Whittaker S, Forom L, Luzzatto L, et al. Lymphatoid granulomatosis: evidence of a clonal T-cell origin and an association with lethal midline granuloma. QuartJ Med 1988; 68: 645-55.

ROUTINE CORONARY ANGIOGRAPHY FOR EFFORT ANGINA

SiR,—The decision to perform coronary angiography in patients with non-disabling effort angina is often based on the results of an exercise test. We question the validity of this approach. In a retrospective analysis 434 patients who had a history of angina at the time of an exercise test which was positive for pain or ST segment changes were followed up for up to 9 years. A conservative management policy was used-ie, coronary angiography was offered to patients with disabling angina with bypass surgery for those with multiple vessel coronary artery obstruction. We found that coronary angiography had been done in

838 CARDIAC CATHETERISATION DATA

COMPARISON OF EXERCISE TEST RESPONSES IN SURVIVORS AND NON-SURVIVORS

Normal values: stroke volume 30-65 ml, systemic vascular resistance 10-20 units, vascular resistance 0 25—1-5 units. To convert mm Hg to kPa multiply by 0,133.

pulmonary

as means (SD). NS =not significant. Stepwise regression analysis based on Cox proportional hazards model indicates that depth of ST segment depression, decision to perform coronary angiography, and age of the patient are significant predictors of survival (when coronary artery score is excluded).

Results

(48%) of patients. These had several indicators of functional impairment on exercise testing: they had had significantly lower heart rates, peak systolic blood pressures and peak work loads than the other 226 patients; they had more angina; and they tended to be After coronary angiography, 125 (60%) younger (all p<0’01). had coronary bypass surgery. The severity of coronary patients artery disease was the major association with bypass surgery (p < 0-001). Depth of ST segment depression and older age were less strongly associated with the decision to advise surgery (p < 005). 51 (11-8%) of the 434 patients died over 6 years. One factor associated with death was non-performance of coronary angiography; another was non-performance of cardiac surgery (though this is not independent), others were severity of coronary artery disease, depth of ST segment depression, and age (table). Despite the greater disability from angina in patients undergoing coronary angiography 95% of these patients were alive at 5 years compared with 86% of patients not having coronary angiography (p < 001). In this study, derived from an exercise test data base, we did not measure other major determinants of survival-ie, left-ventricular 208

narrowing

in the left anterior

descendmg

coronary artery.

Endomyocardial biopsy showed no signs of myocarditis and serological studies for viral infection, blood cultures, and other laboratory tests were negative, apart from a serum angiotensin converting enzyme (ACE) level of 56 U/ml (normal 16-53).3 weeks after the introduction of captopril the patient was still very ill, and cardiac transplantation was considered. In view of his wasted state ( — 16 kg) and panhypopituitarism recombinant-DNA human GH (12 units daily, subcutaneously at 2200 hours) was tried. Within days, he said he felt better; peripheral perfusion increased and mobility improved over the next few months until he could walk indefinitely on the flat and climb two flights of stairs with only minor dyspnoea. In March, 1988, after 3 months of hGH treatment, postural syncopal episodes occurred; clinically he had become sodium

depleted. Frusemide was withdrawn and no further episodes occurred, suggesting that the dose of diuretic, which had previously been essential, had become excessive. The dose of ACE inhibitor was reduced to enalapril 5 mg daily, and hGH was reduced to 4 units daily. The clinical improvement was supported by treadmill exercise testing (Bruce protocol) and measurement of skeletal muscle

function, coronary artery disease risk factors or social class. Nevertheless, we feel that the improved survival in patients

undergoing coronary angiography can be explained by the procedure’s ability to provide a better assessment of risk of dying and by the benefits of bypass surgery in patients selected for this procedure on the basis of angiographic findings. We recommend that all patients with effort angina be evaluated by coronary angiography. St Vincent’s Hospital, Fitzroy, 3065 Victoria, Australia

MICHAEL V. JELINEK NIELS G. BECKER WILLIAM F. RYAN ALEXIA CLEMENS

CARDIAC FAILURE RESPONDING TO GROWTH HORMONE

SIR,--Growth hormone (GH) is a potent anabolic agent. In rats with hypopituitarism GH increases cardiac mass, in keeping with the increase in total body weight,’ but the action of GH on adult cardiac muscle in man is unknown. A 53-year-old man presented in September, 1986, for investigation of Cushing’s syndrome. He had a history of insulin dependent diabetes mellitus. Hypertension had been controlled with atenolol since 1980. Cushing’s disease was confirmed and a total hypophysectomy was done in November, 1986. After major postoperative complications he was discharged on Jan 1, 1987. Heart failure developed in August, 1987, and echocardiography revealed a dilated, diffusely hypokinetic left ventricle and an apical mural thrombus (figure). Despite frusemide and warfarin therapy cardiac function deteriorated. Haemodynamic data are shown in the table. Coronary angiography was normal, apart from a 50%

M-mode

echocardiographic studies. Cushing’s disease diagnosed in September, 1986, cardiac failure developed in August, 1987; institution of captopril in early December, 1987; and hGH treatment between December, 1987, and August, 1988 Normal values: left ventncular (LV) end-diastolic 3-5-5 6 and end-systolic dtmension 2-3-3-7 cm; ejection fraction >40%; septum < 1-2 cm; posterior wall < I’l cm.