193 TREHARNE et al. 14 found
rifampicin more active than erythromycin or oxytetracycline, whilst both groups found sulphonamides, trimethoprim, chloramphenicol, and the penicillins only moderately active. Spectinomycin, metronidazole, and aminoglycosides were inactive. The experience of Kuo, WANG, and GRAYSTON1S was similar: they also noted that rosamicin (a new macrolide antibiotic) was just as active as erythromycin, that different
cycline.
serotypes of C. trachomatis varied in sensitivity to sulphonamides, and that penicillin gave rise to abnormal forms of C. trachomatis. Latency is a well-known phenomenon in C. psittaci infections, and has been suggested as an explanation for some of the clinical findings in oculogenital infections with C. trachomatis.16.17 In vitro, this phenomenon has been seen in C. psittaca’’8 and C. trachomatis.19 Whilst latency was particularly striking with the cell-wall inhibiting antibiotics and the sulphonamides, RIDGWAY et al. 19 found that all antimicrobials tested could induce latency after serial passage. Clearly, caution is required in the interpretation of sensitivity tests with C. trachomatis on single passage. Closely related antimicrobials may have widely differing inhibitory concentrations; thus, results with a single antimicrobial should not be taken to apply to related antimicrobials. Ocular chlamydial infections respond to 5 or 6 weeks’ treatment with tetracycline or rifampicin eye ointment. In 1972, when the role of C. trachomatis in N.G.u. was still in doubt, WILLCOX2o concluded on clinical grounds that ampicillin, nalidixic acid, and novobiocin were no better than placebo; and other compounds of questionable value included sulphonamides, penicillin, chloramphenicol, metronidazole, and spectinomycin. These findings are now borne out by in-vitro tests on C. trachomatis. Tetracyclines, by contrast, have long been known to be effective for N.G.u., and eradication of C. trachomatis coincides with resolution of symptoms.21,22 In patients with gonococcal urethritis ORIEL and his co-workers,9.23,24 showed that coinci dent infection with C. trachomatis was unaffected by gentamicin, ampicillin, and spectinomycin. These patients went on to acquire postgonococcal 14. Treharne, J. D., Day, J., Yeo, C. K., Jones, B. R., Squires, S. in Non Gonococcal Urethritis and Related Infections (edited by D. Hobson and K. K. Holmes); p. 214. American Society for Microbiology, Washington, D.C., 1977. 15. Kuo, C. C., Wang, S. P., Grayston, J. T. Antimicrob. Agents Chemother.
1977, 12, 80. 16. Hannah, L., Dawson, C. R., Briones, O.,
mun. 1968, 101, 43. 17. Richmond, S. J., Hilton,
A. L.,
Clarke,
Thygeson, P., Jawetz,
S. K. Br.
J.
vener.
Dis.
E.
J.
Im-
1972, 48,
437. 18. Hatch, T. P. Infect. Immun. 1975, 12, 211. 19. Ridgway, G. L., Owen, J., Oriel, J. D. Br. J. vener. Dis. (in the press). 20. Willcox, R. R. ibid. 1972, 48, 137. 21. Oriel, J. D., Reeve, P., Nicol, C. S. J. Am. vener. Dis. Ass. 1975, 2, 17. 22. Prentice, M. J., Taylor-Robinson, D., Csonka, G. W. Br. J. vener. Dis. 1976,
52, 269. J. D., Ridgway, G. L., Reeve, P., Beckingham, D. C., Owen, J. J. infect. Dis. 1976, 133, 568. Oriel, J. D., Ridgway, G. L., Tchomouroff, S., Owen, J. Br. J. vener. Dis.
23. Oriel, 24
1977, 53, 226.
urethritis; clearly, when treatment schedules for gonorrhoea are revised, we ought to consider a regimen effective against both Neisseria gonorrhcece and C. trachomatis. Some workers have recommended 3 weeks’ treatment for N.G.U., but it may be unrealistic to expect patients to persevere with medication every six hours for this long, when for much of the time they will be symptom-free. ORIEL et al. 25 found that 250 mg of oxytetracycline sixhourly for 2 weeks gave an 86% cure-rate, as did 500 mg of erythromycin stearate twelve-hourly for 2 weeks. Rifampicin has yet to be evaluated in the treatment of N.G.U. Whether single-dose chemotherapy is feasible (as now practised for gonococcal urethritis) remains to be seen. Treatment with antimicrobials has been employed with some success in the elucidation of other possible agents of N.G.U. HANDSFIELD et a1.26 found that the response of N.G.U. to tetracycline therapy was better in Chlamydia-positive patients than in Chlamydia-negative patients. BowiE and
others" compared sulfisoxazole (predominantly antichlamydial activity), with spectinomycin or
streptomycin (predominantly anti-ureaplasma activity). The results suggested that Chlamydia-negative, ureaplasma-positive urethritis responded better to spectinomycin or streptomycin, indicating a possible role for ureaplasmas in Chlamydia-negative urethritis. Should the female contact be investigated and treated? There is no evidence that contact-tracing as practised with gonorrhoea would reduce the incidence of N.G.U. There is, however, ample evidence of colonisation of female contacts with C. trachomatis,1O,l1 and most of them will be symptomless. Some salpingitis may have a chlamydial aetiology,7 whilst a third of one series of neonatal eye infections were chlamydial. These are two good reasons for treating women with proven C. trachomatis infections, and contacts of men with chlamydial urethritis.
NEONATAL HYPOGLYCÆMIA AND NESIDIOBLASTOSIS NEONATAL hypoglycaemia is usually transient and controllable, by frequent meals or by intravenous glucose. The syndrome is common, especially in small-for-gestational-age infants and in the offspring of mothers with diabetes or erythroblastosis. Persistent hypoglycaemia of infancy and childhood is much less common. It may be familial-and as such the predominant manifestation of multiple endocrine adenomatosis’—or non-familial. 25. Oriel, J. D., Ridgway, G. L., Tchomouroff, S. Scott. med. J. (in the press). 26. Handsfield, H. H., Alexander, E. R., Wang, S. P., Pedersen, A. H. B., Holmes, K. K. J. Am. vener. Dis. Ass. 1976, 2, 5. 27. Bowie, W. R., Floyd, J. F., Miller, Y., Alexander, E. R., Holmes, J., Holmes, K. K. Lancet, 1976. ii, 1276 1. Vance, J. E., Stoll, R. W., Kitabchi, A. F., Buchanan, K. D., Hollander, D., Williams, R. H. Am. J. Med. 1972, 52, 211.
194
Afflicted infants have hyperinsulinaemia, ketotic hypoglycEemia, various endocrine disorders (panhypopituitarism, isolated growth-hormone deficiencies, corticotrophin deficiency, Addison’s disease, hypothyroidism), or hepatic-enzyme deficiency (glycogen-storage diseases, disorders of
gluconeogenesis2). Inappropriate insulin secretion is present in 20-33% of all cases of persistent hypoglycaemia3.4 but diagnosis
is nonetheless difficult. The clinical symptoms are ataxia, somnolence, and coma (or seizures) before meals, starting in most cases before the age of 2. Some workers suggest that hyperinsulinsemic hypoglycarmia is a factor in sudden-infant-death syndrome.5,6 A characteristic finding is a high insulin/glucose ratio in blood, commonly with an associated leucine sensitivity.6,7 The primary goal of management is prevention of the repeated hypoglycxmic attacks, which can damage the brain. Sometimes this can be achieved by frequent feeding combined with cortisol therapy, or by glucose infusions. Alternatively, inappropriate insulin secretion can be controlled with diazoxide or, as lately reported,8 by subcutaneous injections of protamine zinc somatostatin. When medical treatment fails, surgical resection of the pancreas becomes necessary, and even then the patient may need long-term diazoxide therapy.9 Usually at least half the pancreas is resected: except in the rare cases of endocrine adenoma, gross morphological changes are absent. The underlying condition is known as nesidioblastosis-a term coined by Laidlawl° to describe diffuse proliferation of the islet cells, nesidioblasts being defined as "the cells that differentiate out of the duct epithelium to build islets". In the past precise histological diagnosis has been difficult: lack of normal islets, close contact between exocrine ducts and endocrine-cell clusters, and microadenomas have been regarded as hallmarks of the disorder. The process usually affects the entire pancreas. Classification had to await detailed structural analyses of the endocrine pancreas. Kloppel and his colleagues" did morphometric studies of a single case and Svik, Vidnes, and Falkmer9 analysed three cases, using special stains to delineate the A (glucagon) and B (insulin) cells. Useful as these investigations were, they could not properly distinguish the contribution of each individual endocrine cell type. This distinction can be made only by immunocytochemical procedures such as those employed by Heitz et al.,12 who examined seven surgical specimens of pancreas obtained at laparotomy from infants with persistent hyperinsulinaemic hypoglycaemia. They found that four types of endocrine cell invariably participate in the endocrine proliferation-namely, B, A, D (somatostatin), and P.P. (pancreatic polypeptide).Thecells Pagliara, A. S., Karl, I. E., Haymond, M., Kipnis, D. M. Pediatrics, 1973, 82, 365. 3. Kogut, M. D., Blaskovics, M., Donnell, G. N. ibid. 1969, 74, 853. 4. Dr George, A. M., Auerbach, V. H. Am. J. med. Sci. 1960, 240, 792. 5. Polak, J. M., Wigglesworth, J. Lancet, 1976, ii, 570. 6. Cox, J. N., Guelpa, G., Terrapon, M. ibid. p. 739. 7. Cochrane, W. A. Am. J. Dis. Child. 1960, 99, 476. 8. Hirsch, H. J., Loo, S, Evans, N., Crigler, J. F., Jr, Filler, R. M., Gabbay, K. H. New Engl J. Med. 1977, 296, 1323. 9. Søvik, O., Vidnes, J., Falkmer, S. Acta path. microbiol. scand. 1975, 83A, 2.
155. 10. 11.
Laidlaw, G. F. Am. J. Path. 1938, 14, 125. Klöppel, G., Altenähr, E., Reichel, W., Willig, R., Freytag, G. Diabetologia, 1974, 10, 245. 12. Heitz, Ph. U., Klöppel, G., Häcki, W. H., Polak, J. M., Pearse, A. G. E. Diabetes, 1977, 26, 632.
occurred in a ratio which was remarkably constant in the seven cases and for the B and A cells the ratio corresponded to that found in normal islets. On computerised image analysis the endocrine parenchyma proved to be increased almost fivefold in nesidioblastosis compared with age-matched controls, and Sovik et al. concluded that the pathogenesis of the different morphological patterns of endocrine proliferation-multifocal ductuloinsular proliferation, microadenomatosis, focal adenomatosis, and adenoma-are one and the same. Although the prime cause remains unknown, nesidioblastosis evidently represents a failure of the normal mechanisms controlling embryogenesis of the endocrine pancreas and its transition to the normal neonatal and infantile state. A systematic analysis of all hormones known to regulate pancreatic function may reveal the identity of the controlling factor whose excess (stimulation) or lack (failure of inhibition) is responsible for the uncontrolled expansion of the islet cells. WHAT HAPPENS TO CLOTS IN THE LUNGS? WHAT happens to patients with pulmonary emboli who survive to leave hospital? Do they have further pulmonary emboli, with pulmonary hypertension and poor long-term prognosis? There is surprisingly little information about these matters, and two papers from Hall and his co-workers are helpful. In one they report a 9-year follow-up of 72 patients who survived a massive pulmonary embolus after intensive but varied initial treatments.12 patients died within this period but in no case was death due to chronic pulmonary hypertension or to definite recurrence of pulmonary embolism. Of the 56 who were restudied none had clinical signs of pulmonary hypertension, either persistent or of later onset, and only 6% had a possible non-fatal recurrence of pulmonary embolism. Resolution of the embolus was complete in most cases although pulmonary arteriography 2-72 months later showed minor abnormalities in some, as did perfusion lung scans performed up to 8 years after the initial incident. The second paper looks at patients with less severe pulmonary emboli; they had a similar pattern of resolution and recurrence.2 These results are in accord with previous work. None of 58 patients initially studied by Paraskos3 and followed up died of pulmonary embolic disease: survival correlated with the presence or absence of pre-existing heartdisease. When 43 of these patients were restudied after 1-7 years, only 5 had evidence of unresolved embolism and only 1 had chronic cor pulmonale. Definite recurrent pulmonary embolism was recorded in only 1 patient, although 3 others had suspected embolism. Tow and Wagner4 did perfusion lung scans in 69 patients: within 4 months, between half and three-quarters had complete resolution or improvement irrespective of the severity of initial embolism. Others have made similar observations with both lung scans and arteriography, 3-7 1. Hall, R. J. C., Sutton, G. C., Kerr, I. H. Br. Heart J. 1977, 39, 1128. 2. Sutton, G. C., Hall, R. J. C., Kerr, I. H. ibid. p. 1135. 3. Paraskos, J. A., Adelstein, S. J., Smith, R. E., Rickman, F. D., Grossman, W, Dexter, L., Dalen, J. E. New Engl. J. Med. 1973, 289, 55. 4. Tow, D. E., Wagner, H. E. ibid. 1967, 276, 1053. 5. Sautter, R. D., Fletcher, F. W., Emanuel, D. A., Lawton, B. R., Olsen, T. G. J. Am. med. Ass. 1964, 189, 948. 6. Fred, H. L., Axelrad, M. A., Lewis, J. M., Alexander, J. K. ibid. 1966, 196, 1137. 7. Urokinase Pulmonary Embolism Trial. Circulation, 1973, 47, suppl. II, p. 1.
rifampicin more active than erythromycin or oxytetracycline, whilst both groups found sulphonamides, trimethoprim, chloramphenicol, and the penicillins only moderately active. Spectinomycin, metronidazole, and aminoglycosides were inactive. The experience of Kuo, WANG, and GRAYSTON1S was similar: they also noted that rosamicin (a new macrolide antibiotic) was just as active as erythromycin, that different
cycline.
serotypes of C. trachomatis varied in sensitivity to sulphonamides, and that penicillin gave rise to abnormal forms of C. trachomatis. Latency is a well-known phenomenon in C. psittaci infections, and has been suggested as an explanation for some of the clinical findings in oculogenital infections with C. trachomatis.16.17 In vitro, this phenomenon has been seen in C. psittaca’’8 and C. trachomatis.19 Whilst latency was particularly striking with the cell-wall inhibiting antibiotics and the sulphonamides, RIDGWAY et al. 19 found that all antimicrobials tested could induce latency after serial passage. Clearly, caution is required in the interpretation of sensitivity tests with C. trachomatis on single passage. Closely related antimicrobials may have widely differing inhibitory concentrations; thus, results with a single antimicrobial should not be taken to apply to related antimicrobials. Ocular chlamydial infections respond to 5 or 6 weeks’ treatment with tetracycline or rifampicin eye ointment. In 1972, when the role of C. trachomatis in N.G.u. was still in doubt, WILLCOX2o concluded on clinical grounds that ampicillin, nalidixic acid, and novobiocin were no better than placebo; and other compounds of questionable value included sulphonamides, penicillin, chloramphenicol, metronidazole, and spectinomycin. These findings are now borne out by in-vitro tests on C. trachomatis. Tetracyclines, by contrast, have long been known to be effective for N.G.u., and eradication of C. trachomatis coincides with resolution of symptoms.21,22 In patients with gonococcal urethritis ORIEL and his co-workers,9.23,24 showed that coinci dent infection with C. trachomatis was unaffected by gentamicin, ampicillin, and spectinomycin. These patients went on to acquire postgonococcal 14. Treharne, J. D., Day, J., Yeo, C. K., Jones, B. R., Squires, S. in Non Gonococcal Urethritis and Related Infections (edited by D. Hobson and K. K. Holmes); p. 214. American Society for Microbiology, Washington, D.C., 1977. 15. Kuo, C. C., Wang, S. P., Grayston, J. T. Antimicrob. Agents Chemother.
1977, 12, 80. 16. Hannah, L., Dawson, C. R., Briones, O.,
mun. 1968, 101, 43. 17. Richmond, S. J., Hilton,
A. L.,
Clarke,
Thygeson, P., Jawetz,
S. K. Br.
J.
vener.
Dis.
E.
J.
Im-
1972, 48,
437. 18. Hatch, T. P. Infect. Immun. 1975, 12, 211. 19. Ridgway, G. L., Owen, J., Oriel, J. D. Br. J. vener. Dis. (in the press). 20. Willcox, R. R. ibid. 1972, 48, 137. 21. Oriel, J. D., Reeve, P., Nicol, C. S. J. Am. vener. Dis. Ass. 1975, 2, 17. 22. Prentice, M. J., Taylor-Robinson, D., Csonka, G. W. Br. J. vener. Dis. 1976,
52, 269. J. D., Ridgway, G. L., Reeve, P., Beckingham, D. C., Owen, J. J. infect. Dis. 1976, 133, 568. Oriel, J. D., Ridgway, G. L., Tchomouroff, S., Owen, J. Br. J. vener. Dis.
23. Oriel, 24
1977, 53, 226.
urethritis; clearly, when treatment schedules for gonorrhoea are revised, we ought to consider a regimen effective against both Neisseria gonorrhcece and C. trachomatis. Some workers have recommended 3 weeks’ treatment for N.G.U., but it may be unrealistic to expect patients to persevere with medication every six hours for this long, when for much of the time they will be symptom-free. ORIEL et al. 25 found that 250 mg of oxytetracycline sixhourly for 2 weeks gave an 86% cure-rate, as did 500 mg of erythromycin stearate twelve-hourly for 2 weeks. Rifampicin has yet to be evaluated in the treatment of N.G.U. Whether single-dose chemotherapy is feasible (as now practised for gonococcal urethritis) remains to be seen. Treatment with antimicrobials has been employed with some success in the elucidation of other possible agents of N.G.U. HANDSFIELD et a1.26 found that the response of N.G.U. to tetracycline therapy was better in Chlamydia-positive patients than in Chlamydia-negative patients. BowiE and
others" compared sulfisoxazole (predominantly antichlamydial activity), with spectinomycin or
streptomycin (predominantly anti-ureaplasma activity). The results suggested that Chlamydia-negative, ureaplasma-positive urethritis responded better to spectinomycin or streptomycin, indicating a possible role for ureaplasmas in Chlamydia-negative urethritis. Should the female contact be investigated and treated? There is no evidence that contact-tracing as practised with gonorrhoea would reduce the incidence of N.G.U. There is, however, ample evidence of colonisation of female contacts with C. trachomatis,1O,l1 and most of them will be symptomless. Some salpingitis may have a chlamydial aetiology,7 whilst a third of one series of neonatal eye infections were chlamydial. These are two good reasons for treating women with proven C. trachomatis infections, and contacts of men with chlamydial urethritis.
NEONATAL HYPOGLYCÆMIA AND NESIDIOBLASTOSIS NEONATAL hypoglycaemia is usually transient and controllable, by frequent meals or by intravenous glucose. The syndrome is common, especially in small-for-gestational-age infants and in the offspring of mothers with diabetes or erythroblastosis. Persistent hypoglycaemia of infancy and childhood is much less common. It may be familial-and as such the predominant manifestation of multiple endocrine adenomatosis’—or non-familial. 25. Oriel, J. D., Ridgway, G. L., Tchomouroff, S. Scott. med. J. (in the press). 26. Handsfield, H. H., Alexander, E. R., Wang, S. P., Pedersen, A. H. B., Holmes, K. K. J. Am. vener. Dis. Ass. 1976, 2, 5. 27. Bowie, W. R., Floyd, J. F., Miller, Y., Alexander, E. R., Holmes, J., Holmes, K. K. Lancet, 1976. ii, 1276 1. Vance, J. E., Stoll, R. W., Kitabchi, A. F., Buchanan, K. D., Hollander, D., Williams, R. H. Am. J. Med. 1972, 52, 211.
194
Afflicted infants have hyperinsulinaemia, ketotic hypoglycEemia, various endocrine disorders (panhypopituitarism, isolated growth-hormone deficiencies, corticotrophin deficiency, Addison’s disease, hypothyroidism), or hepatic-enzyme deficiency (glycogen-storage diseases, disorders of
gluconeogenesis2). Inappropriate insulin secretion is present in 20-33% of all cases of persistent hypoglycaemia3.4 but diagnosis
is nonetheless difficult. The clinical symptoms are ataxia, somnolence, and coma (or seizures) before meals, starting in most cases before the age of 2. Some workers suggest that hyperinsulinsemic hypoglycarmia is a factor in sudden-infant-death syndrome.5,6 A characteristic finding is a high insulin/glucose ratio in blood, commonly with an associated leucine sensitivity.6,7 The primary goal of management is prevention of the repeated hypoglycxmic attacks, which can damage the brain. Sometimes this can be achieved by frequent feeding combined with cortisol therapy, or by glucose infusions. Alternatively, inappropriate insulin secretion can be controlled with diazoxide or, as lately reported,8 by subcutaneous injections of protamine zinc somatostatin. When medical treatment fails, surgical resection of the pancreas becomes necessary, and even then the patient may need long-term diazoxide therapy.9 Usually at least half the pancreas is resected: except in the rare cases of endocrine adenoma, gross morphological changes are absent. The underlying condition is known as nesidioblastosis-a term coined by Laidlawl° to describe diffuse proliferation of the islet cells, nesidioblasts being defined as "the cells that differentiate out of the duct epithelium to build islets". In the past precise histological diagnosis has been difficult: lack of normal islets, close contact between exocrine ducts and endocrine-cell clusters, and microadenomas have been regarded as hallmarks of the disorder. The process usually affects the entire pancreas. Classification had to await detailed structural analyses of the endocrine pancreas. Kloppel and his colleagues" did morphometric studies of a single case and Svik, Vidnes, and Falkmer9 analysed three cases, using special stains to delineate the A (glucagon) and B (insulin) cells. Useful as these investigations were, they could not properly distinguish the contribution of each individual endocrine cell type. This distinction can be made only by immunocytochemical procedures such as those employed by Heitz et al.,12 who examined seven surgical specimens of pancreas obtained at laparotomy from infants with persistent hyperinsulinaemic hypoglycaemia. They found that four types of endocrine cell invariably participate in the endocrine proliferation-namely, B, A, D (somatostatin), and P.P. (pancreatic polypeptide).Thecells Pagliara, A. S., Karl, I. E., Haymond, M., Kipnis, D. M. Pediatrics, 1973, 82, 365. 3. Kogut, M. D., Blaskovics, M., Donnell, G. N. ibid. 1969, 74, 853. 4. Dr George, A. M., Auerbach, V. H. Am. J. med. Sci. 1960, 240, 792. 5. Polak, J. M., Wigglesworth, J. Lancet, 1976, ii, 570. 6. Cox, J. N., Guelpa, G., Terrapon, M. ibid. p. 739. 7. Cochrane, W. A. Am. J. Dis. Child. 1960, 99, 476. 8. Hirsch, H. J., Loo, S, Evans, N., Crigler, J. F., Jr, Filler, R. M., Gabbay, K. H. New Engl J. Med. 1977, 296, 1323. 9. Søvik, O., Vidnes, J., Falkmer, S. Acta path. microbiol. scand. 1975, 83A, 2.
155. 10. 11.
Laidlaw, G. F. Am. J. Path. 1938, 14, 125. Klöppel, G., Altenähr, E., Reichel, W., Willig, R., Freytag, G. Diabetologia, 1974, 10, 245. 12. Heitz, Ph. U., Klöppel, G., Häcki, W. H., Polak, J. M., Pearse, A. G. E. Diabetes, 1977, 26, 632.
occurred in a ratio which was remarkably constant in the seven cases and for the B and A cells the ratio corresponded to that found in normal islets. On computerised image analysis the endocrine parenchyma proved to be increased almost fivefold in nesidioblastosis compared with age-matched controls, and Sovik et al. concluded that the pathogenesis of the different morphological patterns of endocrine proliferation-multifocal ductuloinsular proliferation, microadenomatosis, focal adenomatosis, and adenoma-are one and the same. Although the prime cause remains unknown, nesidioblastosis evidently represents a failure of the normal mechanisms controlling embryogenesis of the endocrine pancreas and its transition to the normal neonatal and infantile state. A systematic analysis of all hormones known to regulate pancreatic function may reveal the identity of the controlling factor whose excess (stimulation) or lack (failure of inhibition) is responsible for the uncontrolled expansion of the islet cells. WHAT HAPPENS TO CLOTS IN THE LUNGS? WHAT happens to patients with pulmonary emboli who survive to leave hospital? Do they have further pulmonary emboli, with pulmonary hypertension and poor long-term prognosis? There is surprisingly little information about these matters, and two papers from Hall and his co-workers are helpful. In one they report a 9-year follow-up of 72 patients who survived a massive pulmonary embolus after intensive but varied initial treatments.12 patients died within this period but in no case was death due to chronic pulmonary hypertension or to definite recurrence of pulmonary embolism. Of the 56 who were restudied none had clinical signs of pulmonary hypertension, either persistent or of later onset, and only 6% had a possible non-fatal recurrence of pulmonary embolism. Resolution of the embolus was complete in most cases although pulmonary arteriography 2-72 months later showed minor abnormalities in some, as did perfusion lung scans performed up to 8 years after the initial incident. The second paper looks at patients with less severe pulmonary emboli; they had a similar pattern of resolution and recurrence.2 These results are in accord with previous work. None of 58 patients initially studied by Paraskos3 and followed up died of pulmonary embolic disease: survival correlated with the presence or absence of pre-existing heartdisease. When 43 of these patients were restudied after 1-7 years, only 5 had evidence of unresolved embolism and only 1 had chronic cor pulmonale. Definite recurrent pulmonary embolism was recorded in only 1 patient, although 3 others had suspected embolism. Tow and Wagner4 did perfusion lung scans in 69 patients: within 4 months, between half and three-quarters had complete resolution or improvement irrespective of the severity of initial embolism. Others have made similar observations with both lung scans and arteriography, 3-7 1. Hall, R. J. C., Sutton, G. C., Kerr, I. H. Br. Heart J. 1977, 39, 1128. 2. Sutton, G. C., Hall, R. J. C., Kerr, I. H. ibid. p. 1135. 3. Paraskos, J. A., Adelstein, S. J., Smith, R. E., Rickman, F. D., Grossman, W, Dexter, L., Dalen, J. E. New Engl. J. Med. 1973, 289, 55. 4. Tow, D. E., Wagner, H. E. ibid. 1967, 276, 1053. 5. Sautter, R. D., Fletcher, F. W., Emanuel, D. A., Lawton, B. R., Olsen, T. G. J. Am. med. Ass. 1964, 189, 948. 6. Fred, H. L., Axelrad, M. A., Lewis, J. M., Alexander, J. K. ibid. 1966, 196, 1137. 7. Urokinase Pulmonary Embolism Trial. Circulation, 1973, 47, suppl. II, p. 1.
