- Email: [email protected]
ASSOCIATION OF COMPLEMENT ACTIVATION AND ELEVATED PLASMA-C5a WITH ADULT RESPIRATORY DISTRESS SYNDROME
947
ASSOCIATION OF COMPLEMENT ACTIVATION AND ELEVATED PLASMA-C5a WITH ADULT RESPIRATORY DISTRESS SYNDROME
Pathophysiological Relevance and Possible Prognostic Value DALE E. HAMMERSCHMIDT LEONARD D. HUDSON HARRY S.
L. JEAN WEAVER PHILIP R. CRADDOCK
JACOB
Department of Medicine, University of Minnesota, Minneapolis; and Departments of Medicine and Surgery, University of Washington, Seattle, Washington, U.S.A.
Clinical and suggest that
experimental observations aggregation of polymorphonuclear granulocytes (PMN) in response to activated complement (C) might contribute to the genesis of the adult respiratory distress syndrome (ARDS), aggregating PMN causing pulmonary dysfunction by becoming lodged in the lung as leucoemboli. PMN-aggregating activity can be detected in C-activated plasma and reflects C5a levels. In 61 patients at risk for ARDS a strong and highly significant correlation was found between the presence of PMN-aggregating activity in the plasma and the development of ARDS; this correlation was also significant when patients with sepsis were excluded from analysis. In patients followed prospectively detection of elevated C5a seemed to be a useful predictor of ARDS. Since corticosteroids have been shown to inhibit PMN aggregation both in vitro and in vivo, the evidence for a role for PMN aggregation in the genesis of ARDS supports the use of corticosteroids in Summary
this disorder.
<.
Patients and Methods Patients selected for study had a clinically estimated risk of least 0-5 of developing ARDS. Each had suffered major non-thoracic trauma or had fungaemia, gram-negative bacterxmia, or hypotension lasting more than 2 h. Blood samples were taken, and EDTA-anticoagulated plasmas were kept frozen (-70°C) until assay. Plasma samples obtained away from the University of Minnesota were shipped frozen. Negative controls were included in each shipment. The clinical diagnosis of ARDS was made when the followwhere ing features were present: hypoxsemia (PaO/FIO<160, Pa02 is the partial pressure of oxygen in arterial blood and FI02 the fraction of O2 in inspired air), pulmonary interstitial infiltrates on chest radiographs in the absence of clinical pneumonitis, a normal pulmonary-artery wedge pressure, and diminished lung compliance, if measured. PMN aggregation was measured as previously described7 using patients’ plasmas as test aggregants of density-gradientpurified, normal, ABO-compatible PMN. CSa was measured by comparing the resultant aggregation waves with those generated by serial dilutions of a standardised zymosan-activated-plasma (ZAP) preparation (aggregation wave amplitude is a linear function oflogn [C5a] in the sample under study. 10 100 ZAP units were set equal to the [C5a] in reagent ZAP, 50 ZAP units to C5a in a 1:1 dilution of ZAP, &c.11 Negative controls were run with each batch of clinical samples: a sample was considered positive only if its C5a exceeded by 50% that of the highest value obtained on a negative control in the same run. Typical thresholds of detection by these criteria were 0.5-1.5 ZAP units. In a random subset of 12 patients, aggregating activity was shown to be due to C5a by chromatographic estimation of molecular weight and by specific anti-C5 antiserum inhibition of activity.7 As a second highly sensitive assay for complement actiat
Introduction WHEREAS the normal function of the plasma comple(C) system is central to host defences against infecits tion, excessive or inappropriate activation may harm the host, as in immune arthritis or nephritis. Several clinical and experimental parallels have suggested that the deleterious consequences of complement activation may also operate in the adult respiratory distress syndrome (ARDS), or "shock lung". First, the microvascular stasis of granulocytes is an early pathological hallmark of ARDS.1 This same leucostasis can be produced in animals by intravenous infusion of activated C5a;2-4 in fact, long-term infusion results in a necrotising vasculitis which may be an analogue of early severe ARDS (T. K. Bowers, N. Ratliff, A. Ozalins, and H. S. Jacob, unpublished observations). Secondly, among the most common clinical contexts for the development of ARDS are sepsis, pancreatitis, and severe trauma’-all states potentially accompanied by massive complement activation. Thirdly, animals depleted of either granulocytes or activatable C are relatively resistant to the development of "shock lung" or ment
irreversible shock.5.6
microvascular endothelium might then be wrought by stagnant polymorphonuclear granulocytes (PMN); such damage might then initiate or amplify ARDS. A more detailed mechanism causing leucostasis and endothelial damage is suggested by other studies from our laboratory. In studying the complement-mediated leucostasis and pulmonary dysfunction which attend hxmodialysis, we discovered that PMN will aggregate in response to the generation of C5a and that such aggregation may be observed and measured with essentially the same nephelometric techniques as are used to study platelet aggregation.7 Thus, when C is activated, PMN may aggregate and then lodge in the lung as leucoemboli. Further, we have shown8 that C-stimulated PMN may damage cultured endothelial cells in vitro, in large measure by the release of toxic oxygen species. In addition to suggesting a possible pathophysiological mechanism for ARDS, the aggregation of granulocytes in response to C5a also provided a simple quantitative assay for C5a in clinical samples which was a more sensitive detector of C activation than the assays in general clinical use.9 As a first clinical test of our pathophysiological hypothesis, we examined for C5a plasmas from patients at risk for ARDS and looked for a correlation between this and other evidence of C activation and the development of ARDS. In the present study such a correlation was in fact found, and study of serial samples from several of the patients suggested that evidence of complement activation may be a clinically useful predictor of the development of ARDS.
,
From these considerations, we hypothesised that longterm or excessive complement activation in consequence of an underlying disease process might result in pulmonary leucostasis, and secondary damage to pulmonary
948 TABLE I-RISK FACTORS FOR DEVELOPMENT OF ARDS
TABLE HI——CORRELATION OF POSITIVE
PLASMA-C5a
ASSAYS
WITH DEVELOPMENT OF ARDS IN PATIENTS FOLLOWED
PROSPECTIVELY
were also examined by for the presence of C3 conversion
vation, the plasmas
immunoelectro-
products, with methods previously described. 12 All assays were done by staff unaware of the clinical status of the patients under study. ARDS or its absence was diagnosed by physicians unaware of the results of complement studies.
phoresis
clinically.
Results Of 61 patients studied, 33 developed the syndrome of ARDS and 28 did not; the observed risk ratio of 0.54 is close to that sought for inclusion in the study. The risk factors predisposing to ARDS are shown in table I. A correlation was observed between complement activation and the development of ARDS. A patient was considered C5a-positive if at least one CSa assay was unequivocally positive on the day of diagnosis of ARDS or in the preceding 72 h (table n). 31 of the 33 patients ultimately developing ARDS had positive assays, whereas only 5 of the 28 patients not developing ARDS were positive (x2=363; p«0-00001). This correlation persisted even if septic patients were excluded from analysis (table n). In that subset, 18 of 19 patients developing ARDS had positive C5a assays; only 3 positive assays were observed among 26 patients who did not develop ARDS (-31-9; p< <0.00001). Both C5a and C3 conversion assays were done in 37 of the patients; addition of the C3 conversion assay only very slightly improved the results. In the 16 of these patients who developed ARDS, only 1 had a positive C3 conversion assay in association with negative C5a assays; in contrast, 6 patients had positive C5a assays without demonstrable C3 conversion. In patients not developing ARDS, the C3 conversion and C5a assays were concordant (both positive in 3, both negative in 18). The C5a assay was therefore the more sensitive detector of C activation: the C3 conversion assay eliminated only 1 false-negative result and had no effect upon true-negative or false-positive results. .31 patients had serial plasmas collected over several days, beginning with the time that high risk for ARDS was recognised. Unfortunately, only 6 of the ARDS paTABLE II-ASSOCIATION OF POSITIVE PLASMA DEVELOPMENT OF ARDS
tients were included in this group. In this prospectively followed group, the correlation of C5a positivity with ARDS was even more pronounced (table in) (=31; 5 of the 6 positive patients had positive C5a p<<0-001). h 8 or more before ARDS could be diagnosed assays
C5a
ASSAYS WITH
Discussion Clinical and experimental parallels suggested that interaction between granulocytes and the plasma-complement cascade might play a part in the pathogenesis of ARDS or "shock lung". Our interest in this possibility was focused by the observation that PMN aggregate in response to complement activation’-an observation which led us to consider the previously unsuspected pathophysiological mechanism of leucoembolisation in immune organ damage.13-15 The feasibility of this postu-
lated mechanism was then supported by the observation of C5a-induced leucoembolisation by intravital microscopy of rat mesenteriesl6 and of endothelial damage by C 5 a-stimulated PMN in tissue culture. did measurement of PMN aggregation proinsight into possible immune pathogenesis: it proved to be a uniquely sensitive detector of C activation, providing a good estimate of C5a in clinical samples.9 This complement assay therefore seemed the ideal tool for a first clinical test of our hypotheses. We therefore sought correlations between abnormal C5a in plasma and the development of ARDS in patients judged clinically at risk for that complication. Measurable C5a is normally absent from plasma. However, slight C activation may occur as an artefact in sample handling,l’ and technical difficulties may arise with the assay. Negative controls were therefore included with each batch of samples, and a threshold of reliable detection was estimated for each batch (usually about 1.5 ZAP units-an arbitrary scale in use in our laboratory until a molecular concentration equivalent scale can be devised). Indeed, when patients with ARDS were studied, 31 of 33 had abnormal plasma C5a levels on the day of diagnosis or during the preceding 72 h. In contrast, only 5 of 28 high-risk patients who did not develop ARDS had excess plasma C5a, a difference which was highly statistically significant (table 11). One of the commonest predisposing factors in this series-and the one most obviously associated with complement activation-was bacterwmic or fungxmic sepsis. A large proportion of patients with sepsis (14 of 16, table i) developed ARDS, so it was possible that our observed correlation was an Not
vide also
only
a new
949 more than that C is activated in We therefore re-evaluated our results after sepsis. with sepsis (table n); a similar highly excluding patients was observed. There was no sigsignificant relationship nificant difference between the distributions observed in patients without sepsis and all patients (<0-3; p>0-2). In several of the patients, more complete C profiles were available, including haemolytic C3 and total C levels. Although there were too few such patients for meaningful analysis, at least 6 ARDS patients showed C3 conversion and/or elevated plasma C5a while maintaining normal CHso and C3Hso levels. Thus, the association described here might have escaped detection, had less sensitive assays of C activation, been used. While these observations support our pathophysiological hypothesis that C activation is important in the genesis of ARDS, they also suggest the even more exciting possibility that evidence of C activation may be a useful predictor of ARDS in susceptible patients. Thus, of 6 ARDS patients who had prospective sampling done from the time of admission to hospital, 5 developed elevated plasma C5a levels 8-72 h before the clinical recognition of ARDS. We have reported that very high concentrations of corticosteroids, (especially methylprednisolone) will inhibit granulocyte aggregation both in vitro and in vivo. 15 Similar concentrations have been advocated for use in ARDS.18 In supporting the possible role of PMN activation in the genesis of ARDS, the current study also supports the appropriateness of corticosteroid therapy by suggesting a mechanism for the beneficial effect of such agents in ARDS. We are beginning a study in which all patients at risk of ARDS will be followed prospectively, and prophylactic measures (positive end-expiratory pressure ventilation and corticosteroids) will be used at random. From such a study we hope to establish whether C-activation studies will predict ARDS or identify a group of patients likely to benefit from early therapeutic intervention. We thank Nan Wilson and Carol Lammi for technical assistance, Dr Enc Overland and Dr S. Schonfield for providing clinical samples, and the residents and other house-staff for obtaining plasma samples. This work is supported by National Institutes of Health grants GM 24990, AM 15730, HL 19725, CA 15627, and HL 07062. P. R. C. is the recipient of a research career development award from the N.I.H. (1 K04HL 00479). D. E. H. is the recipient of a young investigator research award (1 R23-HL 25043) from the N.I.H. Requests for reprints should be addressed to D. E. H., Box 480 Mayo, University of Minnesota Hospital, Minneapolis, Minnesota 55455, U.S.A.
artefact meaning little
PRENATAL DIAGNOSIS OF EPIDERMOLYSIS BULLOSA LETALIS C. H. RODECK
Department of Obstetrics and Gynæcology, King’s College Hospital Medical School, London SE5 8RX R. A. Electron Microscopy
J. EADY
Unit, Institute of Dermatology, London E9
C. M. GOSDEN
’
REFERENCES 1. Sandritter
WC, Mittermayer C, Riede UN, Freudenberg N, Grimm H. Das Schocklungen-Syndrom (ein allgemeiner Uberblick.) Beitr Pathol 1978;
162: 7-23. 2. Craddock PR, Fehr
J, Dalmasso AP, Brigham KL, Jacob HS. Hemodialysis leukopenia: Pulmonary vascular leukostasis resulting from complement activation by dialyzer cellophane membranes. J Clin Invest 1977; 99:
879-88. 3. Craddock PR, Fehr J, Brigham KL, Kronenberg RS, Jacob HS. Complement and leukocyte-mediated pulmonary dysfunction in hemodialysis. N Engl J Med 1977; 296: 769-74. 4. Fehr J, Jacob HS. In-vitro granulocyte adherence and in-vivo margination: Two associated complement-dependent functions J Exp Med 1977; 146: 641-52. 5. Pingleton WW, Coalson
JJ, Guenter CA. Significance of leukocytes m endotoxic shock. Exp Mol Pathol 1978; 22: 183-94. 6. Hosea SW, Hammer CH, Frank M. The role of complement in the respiratory distress syndrome. Clin Res 1978; 26: 397A.
MRC Clinical and Population Cytogenetics Unit, Western General Hospital, Edinburgh
Summary roscopy of a direct vision
diagnosis of epidermolysis was made by electron micbiopsy specimen of fetal skin taken under by fetoscopy at 18 weeks’ gestation. The The prenatal bullosa letalis
confirmed after termination of pregnancy. were found in the number, size, and scanning-electron-microscope appearance of amniotic-fluid cells. These techniques have considerable potential for rapid diagnostic studies in a number of other conditions.
diagnosis
was
Abnormalities
Introduction EPIDERMOLYSis bullosa letalisl (EBL) is a severe, usually fatal, skin disorder characterised by blistering and often associated with erosions of the mucous membranes of the mouth and other parts of the alimentary tract. The defect occurs in the junction between the epidermis and dermis; transmission electron microscopy
PR, Hammerschmidt DE, White JG, Dalmasso AP, Jacob HS. - Complement (C5a)-induced granulocyte aggregation in vitro. A possible mechanism of complement-medicated leukostasis and leukopenia. J Clin
7. Craddock
Invest 1977; 60:260. 8. Sacks T, Moldow CF, Craddock PR, Bowers TK, Jacob HS. Oxygen radicals mediate endothelial damage by complement-stimulated granulocytes: An in-vitro model of immune vascular damage. Clin Invest 1978; 61: 1161-67. 9. Hammerschmidt DE, Bowers TK, Lammi-Keefe CJ, Jacob HS, Craddock PR. Granulocyte aggregometry: A sensitive technique for the detection of C5a and complement activation. Blood (m press). 10. Craddock PR, White JG, Jacobs HS. Potentiation of complement (C5a)induced granulocyte aggregation by cytochalasin B. J Lab Clin Med 1978; 91: 490-99. 11. Schmeling DJ, Peterson PK, Hammerschmidt DE, et al. Chemotaxigenesis by cell surface components of Staphylococcus aureus. Infect Immunity
1979; 26: 57-63. 12. Hammerschmidt DE, Craddock PR, McCullough JJ, Kronenberg RS, Dalmasso AP, Jacob HS. Complement activation and pulmonary leukostasis during nylon fiber filtration leukapheresis. Blood 1978; 51: 721-30. 13. Jacob HS. Granulocyte-complement interaction: A beneficial anti-microbial mechanism that can cause disease. Arch Intern Med 1978; 138: 461-68. 14. Greenberg CS, Hammerschmidt DE, Craddock PR, Jacob HS. Atheroma cholesterol activates complement and aggregates granulocytes: Possible role in ischemic manifestations of atherosclerosis. Trans Ass Am Physns
(in press). 15. Hammerschmidt DE, White JG, Craddock PR, Jacob HS. Corticosteroids inhibit complement induced granulocyte aggregation: A possible mechanism for their efficacy in shock states. J Clin Invest 1979; 63: 798-803. 16. Hammerschmidt DE, Harris PD, Wayland JH. Jacob HS. Intravascular granulocyte aggregation in live animals: A complement-mediated mechanism of ischemia. Blood 1978; 52 (suppl. 1): 125. 17. Hammerschmidt DE, Wilson N. Artefactual complement activation by blood-drawing apparatus: A clinical and investigational caveat. Am J Clin Pathol (in press). 18. Slader A. Methylprednisolone. Pharmacologic doses in shock lung syndrome. J Thorac Cardiovasc Surg 1976; 21: 800-05.
ASSOCIATION OF COMPLEMENT ACTIVATION AND ELEVATED PLASMA-C5a WITH ADULT RESPIRATORY DISTRESS SYNDROME
Pathophysiological Relevance and Possible Prognostic Value DALE E. HAMMERSCHMIDT LEONARD D. HUDSON HARRY S.
L. JEAN WEAVER PHILIP R. CRADDOCK
JACOB
Department of Medicine, University of Minnesota, Minneapolis; and Departments of Medicine and Surgery, University of Washington, Seattle, Washington, U.S.A.
Clinical and suggest that
experimental observations aggregation of polymorphonuclear granulocytes (PMN) in response to activated complement (C) might contribute to the genesis of the adult respiratory distress syndrome (ARDS), aggregating PMN causing pulmonary dysfunction by becoming lodged in the lung as leucoemboli. PMN-aggregating activity can be detected in C-activated plasma and reflects C5a levels. In 61 patients at risk for ARDS a strong and highly significant correlation was found between the presence of PMN-aggregating activity in the plasma and the development of ARDS; this correlation was also significant when patients with sepsis were excluded from analysis. In patients followed prospectively detection of elevated C5a seemed to be a useful predictor of ARDS. Since corticosteroids have been shown to inhibit PMN aggregation both in vitro and in vivo, the evidence for a role for PMN aggregation in the genesis of ARDS supports the use of corticosteroids in Summary
this disorder.
<.
Patients and Methods Patients selected for study had a clinically estimated risk of least 0-5 of developing ARDS. Each had suffered major non-thoracic trauma or had fungaemia, gram-negative bacterxmia, or hypotension lasting more than 2 h. Blood samples were taken, and EDTA-anticoagulated plasmas were kept frozen (-70°C) until assay. Plasma samples obtained away from the University of Minnesota were shipped frozen. Negative controls were included in each shipment. The clinical diagnosis of ARDS was made when the followwhere ing features were present: hypoxsemia (PaO/FIO<160, Pa02 is the partial pressure of oxygen in arterial blood and FI02 the fraction of O2 in inspired air), pulmonary interstitial infiltrates on chest radiographs in the absence of clinical pneumonitis, a normal pulmonary-artery wedge pressure, and diminished lung compliance, if measured. PMN aggregation was measured as previously described7 using patients’ plasmas as test aggregants of density-gradientpurified, normal, ABO-compatible PMN. CSa was measured by comparing the resultant aggregation waves with those generated by serial dilutions of a standardised zymosan-activated-plasma (ZAP) preparation (aggregation wave amplitude is a linear function oflogn [C5a] in the sample under study. 10 100 ZAP units were set equal to the [C5a] in reagent ZAP, 50 ZAP units to C5a in a 1:1 dilution of ZAP, &c.11 Negative controls were run with each batch of clinical samples: a sample was considered positive only if its C5a exceeded by 50% that of the highest value obtained on a negative control in the same run. Typical thresholds of detection by these criteria were 0.5-1.5 ZAP units. In a random subset of 12 patients, aggregating activity was shown to be due to C5a by chromatographic estimation of molecular weight and by specific anti-C5 antiserum inhibition of activity.7 As a second highly sensitive assay for complement actiat
Introduction WHEREAS the normal function of the plasma comple(C) system is central to host defences against infecits tion, excessive or inappropriate activation may harm the host, as in immune arthritis or nephritis. Several clinical and experimental parallels have suggested that the deleterious consequences of complement activation may also operate in the adult respiratory distress syndrome (ARDS), or "shock lung". First, the microvascular stasis of granulocytes is an early pathological hallmark of ARDS.1 This same leucostasis can be produced in animals by intravenous infusion of activated C5a;2-4 in fact, long-term infusion results in a necrotising vasculitis which may be an analogue of early severe ARDS (T. K. Bowers, N. Ratliff, A. Ozalins, and H. S. Jacob, unpublished observations). Secondly, among the most common clinical contexts for the development of ARDS are sepsis, pancreatitis, and severe trauma’-all states potentially accompanied by massive complement activation. Thirdly, animals depleted of either granulocytes or activatable C are relatively resistant to the development of "shock lung" or ment
irreversible shock.5.6
microvascular endothelium might then be wrought by stagnant polymorphonuclear granulocytes (PMN); such damage might then initiate or amplify ARDS. A more detailed mechanism causing leucostasis and endothelial damage is suggested by other studies from our laboratory. In studying the complement-mediated leucostasis and pulmonary dysfunction which attend hxmodialysis, we discovered that PMN will aggregate in response to the generation of C5a and that such aggregation may be observed and measured with essentially the same nephelometric techniques as are used to study platelet aggregation.7 Thus, when C is activated, PMN may aggregate and then lodge in the lung as leucoemboli. Further, we have shown8 that C-stimulated PMN may damage cultured endothelial cells in vitro, in large measure by the release of toxic oxygen species. In addition to suggesting a possible pathophysiological mechanism for ARDS, the aggregation of granulocytes in response to C5a also provided a simple quantitative assay for C5a in clinical samples which was a more sensitive detector of C activation than the assays in general clinical use.9 As a first clinical test of our pathophysiological hypothesis, we examined for C5a plasmas from patients at risk for ARDS and looked for a correlation between this and other evidence of C activation and the development of ARDS. In the present study such a correlation was in fact found, and study of serial samples from several of the patients suggested that evidence of complement activation may be a clinically useful predictor of the development of ARDS.
,
From these considerations, we hypothesised that longterm or excessive complement activation in consequence of an underlying disease process might result in pulmonary leucostasis, and secondary damage to pulmonary
948 TABLE I-RISK FACTORS FOR DEVELOPMENT OF ARDS
TABLE HI——CORRELATION OF POSITIVE
PLASMA-C5a
ASSAYS
WITH DEVELOPMENT OF ARDS IN PATIENTS FOLLOWED
PROSPECTIVELY
were also examined by for the presence of C3 conversion
vation, the plasmas
immunoelectro-
products, with methods previously described. 12 All assays were done by staff unaware of the clinical status of the patients under study. ARDS or its absence was diagnosed by physicians unaware of the results of complement studies.
phoresis
clinically.
Results Of 61 patients studied, 33 developed the syndrome of ARDS and 28 did not; the observed risk ratio of 0.54 is close to that sought for inclusion in the study. The risk factors predisposing to ARDS are shown in table I. A correlation was observed between complement activation and the development of ARDS. A patient was considered C5a-positive if at least one CSa assay was unequivocally positive on the day of diagnosis of ARDS or in the preceding 72 h (table n). 31 of the 33 patients ultimately developing ARDS had positive assays, whereas only 5 of the 28 patients not developing ARDS were positive (x2=363; p«0-00001). This correlation persisted even if septic patients were excluded from analysis (table n). In that subset, 18 of 19 patients developing ARDS had positive C5a assays; only 3 positive assays were observed among 26 patients who did not develop ARDS (-31-9; p< <0.00001). Both C5a and C3 conversion assays were done in 37 of the patients; addition of the C3 conversion assay only very slightly improved the results. In the 16 of these patients who developed ARDS, only 1 had a positive C3 conversion assay in association with negative C5a assays; in contrast, 6 patients had positive C5a assays without demonstrable C3 conversion. In patients not developing ARDS, the C3 conversion and C5a assays were concordant (both positive in 3, both negative in 18). The C5a assay was therefore the more sensitive detector of C activation: the C3 conversion assay eliminated only 1 false-negative result and had no effect upon true-negative or false-positive results. .31 patients had serial plasmas collected over several days, beginning with the time that high risk for ARDS was recognised. Unfortunately, only 6 of the ARDS paTABLE II-ASSOCIATION OF POSITIVE PLASMA DEVELOPMENT OF ARDS
tients were included in this group. In this prospectively followed group, the correlation of C5a positivity with ARDS was even more pronounced (table in) (=31; 5 of the 6 positive patients had positive C5a p<<0-001). h 8 or more before ARDS could be diagnosed assays
C5a
ASSAYS WITH
Discussion Clinical and experimental parallels suggested that interaction between granulocytes and the plasma-complement cascade might play a part in the pathogenesis of ARDS or "shock lung". Our interest in this possibility was focused by the observation that PMN aggregate in response to complement activation’-an observation which led us to consider the previously unsuspected pathophysiological mechanism of leucoembolisation in immune organ damage.13-15 The feasibility of this postu-
lated mechanism was then supported by the observation of C5a-induced leucoembolisation by intravital microscopy of rat mesenteriesl6 and of endothelial damage by C 5 a-stimulated PMN in tissue culture. did measurement of PMN aggregation proinsight into possible immune pathogenesis: it proved to be a uniquely sensitive detector of C activation, providing a good estimate of C5a in clinical samples.9 This complement assay therefore seemed the ideal tool for a first clinical test of our hypotheses. We therefore sought correlations between abnormal C5a in plasma and the development of ARDS in patients judged clinically at risk for that complication. Measurable C5a is normally absent from plasma. However, slight C activation may occur as an artefact in sample handling,l’ and technical difficulties may arise with the assay. Negative controls were therefore included with each batch of samples, and a threshold of reliable detection was estimated for each batch (usually about 1.5 ZAP units-an arbitrary scale in use in our laboratory until a molecular concentration equivalent scale can be devised). Indeed, when patients with ARDS were studied, 31 of 33 had abnormal plasma C5a levels on the day of diagnosis or during the preceding 72 h. In contrast, only 5 of 28 high-risk patients who did not develop ARDS had excess plasma C5a, a difference which was highly statistically significant (table 11). One of the commonest predisposing factors in this series-and the one most obviously associated with complement activation-was bacterwmic or fungxmic sepsis. A large proportion of patients with sepsis (14 of 16, table i) developed ARDS, so it was possible that our observed correlation was an Not
vide also
only
a new
949 more than that C is activated in We therefore re-evaluated our results after sepsis. with sepsis (table n); a similar highly excluding patients was observed. There was no sigsignificant relationship nificant difference between the distributions observed in patients without sepsis and all patients (<0-3; p>0-2). In several of the patients, more complete C profiles were available, including haemolytic C3 and total C levels. Although there were too few such patients for meaningful analysis, at least 6 ARDS patients showed C3 conversion and/or elevated plasma C5a while maintaining normal CHso and C3Hso levels. Thus, the association described here might have escaped detection, had less sensitive assays of C activation, been used. While these observations support our pathophysiological hypothesis that C activation is important in the genesis of ARDS, they also suggest the even more exciting possibility that evidence of C activation may be a useful predictor of ARDS in susceptible patients. Thus, of 6 ARDS patients who had prospective sampling done from the time of admission to hospital, 5 developed elevated plasma C5a levels 8-72 h before the clinical recognition of ARDS. We have reported that very high concentrations of corticosteroids, (especially methylprednisolone) will inhibit granulocyte aggregation both in vitro and in vivo. 15 Similar concentrations have been advocated for use in ARDS.18 In supporting the possible role of PMN activation in the genesis of ARDS, the current study also supports the appropriateness of corticosteroid therapy by suggesting a mechanism for the beneficial effect of such agents in ARDS. We are beginning a study in which all patients at risk of ARDS will be followed prospectively, and prophylactic measures (positive end-expiratory pressure ventilation and corticosteroids) will be used at random. From such a study we hope to establish whether C-activation studies will predict ARDS or identify a group of patients likely to benefit from early therapeutic intervention. We thank Nan Wilson and Carol Lammi for technical assistance, Dr Enc Overland and Dr S. Schonfield for providing clinical samples, and the residents and other house-staff for obtaining plasma samples. This work is supported by National Institutes of Health grants GM 24990, AM 15730, HL 19725, CA 15627, and HL 07062. P. R. C. is the recipient of a research career development award from the N.I.H. (1 K04HL 00479). D. E. H. is the recipient of a young investigator research award (1 R23-HL 25043) from the N.I.H. Requests for reprints should be addressed to D. E. H., Box 480 Mayo, University of Minnesota Hospital, Minneapolis, Minnesota 55455, U.S.A.
artefact meaning little
PRENATAL DIAGNOSIS OF EPIDERMOLYSIS BULLOSA LETALIS C. H. RODECK
Department of Obstetrics and Gynæcology, King’s College Hospital Medical School, London SE5 8RX R. A. Electron Microscopy
J. EADY
Unit, Institute of Dermatology, London E9
C. M. GOSDEN
’
REFERENCES 1. Sandritter
WC, Mittermayer C, Riede UN, Freudenberg N, Grimm H. Das Schocklungen-Syndrom (ein allgemeiner Uberblick.) Beitr Pathol 1978;
162: 7-23. 2. Craddock PR, Fehr
J, Dalmasso AP, Brigham KL, Jacob HS. Hemodialysis leukopenia: Pulmonary vascular leukostasis resulting from complement activation by dialyzer cellophane membranes. J Clin Invest 1977; 99:
879-88. 3. Craddock PR, Fehr J, Brigham KL, Kronenberg RS, Jacob HS. Complement and leukocyte-mediated pulmonary dysfunction in hemodialysis. N Engl J Med 1977; 296: 769-74. 4. Fehr J, Jacob HS. In-vitro granulocyte adherence and in-vivo margination: Two associated complement-dependent functions J Exp Med 1977; 146: 641-52. 5. Pingleton WW, Coalson
JJ, Guenter CA. Significance of leukocytes m endotoxic shock. Exp Mol Pathol 1978; 22: 183-94. 6. Hosea SW, Hammer CH, Frank M. The role of complement in the respiratory distress syndrome. Clin Res 1978; 26: 397A.
MRC Clinical and Population Cytogenetics Unit, Western General Hospital, Edinburgh
Summary roscopy of a direct vision
diagnosis of epidermolysis was made by electron micbiopsy specimen of fetal skin taken under by fetoscopy at 18 weeks’ gestation. The The prenatal bullosa letalis
confirmed after termination of pregnancy. were found in the number, size, and scanning-electron-microscope appearance of amniotic-fluid cells. These techniques have considerable potential for rapid diagnostic studies in a number of other conditions.
diagnosis
was
Abnormalities
Introduction EPIDERMOLYSis bullosa letalisl (EBL) is a severe, usually fatal, skin disorder characterised by blistering and often associated with erosions of the mucous membranes of the mouth and other parts of the alimentary tract. The defect occurs in the junction between the epidermis and dermis; transmission electron microscopy
PR, Hammerschmidt DE, White JG, Dalmasso AP, Jacob HS. - Complement (C5a)-induced granulocyte aggregation in vitro. A possible mechanism of complement-medicated leukostasis and leukopenia. J Clin
7. Craddock
Invest 1977; 60:260. 8. Sacks T, Moldow CF, Craddock PR, Bowers TK, Jacob HS. Oxygen radicals mediate endothelial damage by complement-stimulated granulocytes: An in-vitro model of immune vascular damage. Clin Invest 1978; 61: 1161-67. 9. Hammerschmidt DE, Bowers TK, Lammi-Keefe CJ, Jacob HS, Craddock PR. Granulocyte aggregometry: A sensitive technique for the detection of C5a and complement activation. Blood (m press). 10. Craddock PR, White JG, Jacobs HS. Potentiation of complement (C5a)induced granulocyte aggregation by cytochalasin B. J Lab Clin Med 1978; 91: 490-99. 11. Schmeling DJ, Peterson PK, Hammerschmidt DE, et al. Chemotaxigenesis by cell surface components of Staphylococcus aureus. Infect Immunity
1979; 26: 57-63. 12. Hammerschmidt DE, Craddock PR, McCullough JJ, Kronenberg RS, Dalmasso AP, Jacob HS. Complement activation and pulmonary leukostasis during nylon fiber filtration leukapheresis. Blood 1978; 51: 721-30. 13. Jacob HS. Granulocyte-complement interaction: A beneficial anti-microbial mechanism that can cause disease. Arch Intern Med 1978; 138: 461-68. 14. Greenberg CS, Hammerschmidt DE, Craddock PR, Jacob HS. Atheroma cholesterol activates complement and aggregates granulocytes: Possible role in ischemic manifestations of atherosclerosis. Trans Ass Am Physns
(in press). 15. Hammerschmidt DE, White JG, Craddock PR, Jacob HS. Corticosteroids inhibit complement induced granulocyte aggregation: A possible mechanism for their efficacy in shock states. J Clin Invest 1979; 63: 798-803. 16. Hammerschmidt DE, Harris PD, Wayland JH. Jacob HS. Intravascular granulocyte aggregation in live animals: A complement-mediated mechanism of ischemia. Blood 1978; 52 (suppl. 1): 125. 17. Hammerschmidt DE, Wilson N. Artefactual complement activation by blood-drawing apparatus: A clinical and investigational caveat. Am J Clin Pathol (in press). 18. Slader A. Methylprednisolone. Pharmacologic doses in shock lung syndrome. J Thorac Cardiovasc Surg 1976; 21: 800-05.