Acute respiratory disorder, rhinoconjunctivitis and fever associated with the pyrolysis of polyurethane derived from

Acute respiratory disorder, rhinoconjunctivitis and fever associated with the pyrolysis of polyurethane derived from diphenylmethane diisocyanate. Scand J Work Environ Health 1994;20:216-22. OBJECTIVES - A case is described of complex reactions associated with exposure to diphenylmeth ane diisocyanate (MDI), with some immunologic observations. METHODS - Medical history, clinical examinations, and analyses of immunologic parameters and the 4,4'-MDI-related amine 4,4'-diaminodiphenylmethane (MDA) in hydrolyzed serum and urine were used. RESULTS - The patient, a mechanic whose medical history suggested repeated attacks of a work-re lated pulmonary or systemic disease, was examined because of acute respiratory disorder, rhinocon junctivitis, and a late systemic reaction after exposure to polyurethane pyrolysis product s, including 4,4'-MDI (air level 15 Ilg ' rrr'). Spirometry showed a partly reversible obstructive dysfunction, and a skin-prick test was positive versus isocyanates conjugated with human serum album in (HSA). MDA was detected in hydrolyzed serum (5.6 ng . ml) and urine (1.6 ug . g creatinine."). In serum, there were specific immmunogiobulin (Ig) G (IgGI and IgG4) and IgE antibodies to 4,4'-MDI-HSA and other isocyanates (phenylisocyanate, toluene diisocyanate, p-toluene monoisocyanate, hexamethylene diiso cyanate) conjugated with HSA, a very high total IgE, a raised total IgG, and moderate neutrophilia and eosinophilia. The specific antibodies declined, but were still increased five years later. Further more, the values of circulating immune complexes were high. In vitro, the circulating immune com plexes in serum increased after the addition of 4,4'-MDI-HSA. The patient had anti-Clq antibodies, which probably accounted for part of the circulating immune compl exes. CONCLUSIONS - The reactions associated with MDI exposure (in combination with exposure to py rolysis products) had features compatible with immediate hypersensitivity and with a complement mediated immune-complex reaction.

1976, he started to manufacture, assemble, and repair conveyer belts [made of PU , polyvinyl chloride (P VC ), polyester, or natural fibers] in his workshop and in different fa cto ries . He j oined PU or PVC belts u sing hi gh-frequency welding o r a homemade tool, which heated th e material electrically to 150-170°C. He al so imprinted belts with another homem ade tool , a metal roller, while blowing hot air (350-600°C) o n the su rface o f th e belts .
The patient h ad never had a ny a lle rgic m an ifestations. He w as a sm o k er (a bo ut 30 cigarettes daily ). In 1970 he had unilateral o ptic neuritis . In Janu ary 1984 he w as hospitalized, after three we eks of coug hing with hemoptysis, because of a pulmonary in filtrate in the ri ght upper lobe. The infiltrate o nly g ra d ua lly sub si ded after he was put on antibiotics. Bronchoscopy re vealed generalized bronchitis. Two m onths later, he had a flu-like illness. In Jul y 1986 he spe nt 24 h in the hospital under the diagnosi s "virosis," bec au se of repeated attacks of m yalgia , fever , and shivering dur ing the preceding fortnight. His temp erature was 38.7°C at admiss ion, (next morn ing 37.0°C). He had been wor king with isocyanate glue for about a week, and, on the day of hospitalization, had been joining PU conveye r belts.
In the beginn ing of November 1986, he had been imprinting on a PU-impregnated conveyer belt ( 19 x 0.4 m; PU surfac e 0.8 mm thick ) for 5 h in a bakery. After work, he suffered from chest tightn ess. Between 1300 and 1500 in the afternoon on the 19th of Novembe r, he had again imprinted on the belt. He used no respiratory protective equipment. Fifteen minut es later he was struc k with heavy respiratory distress, which lasted for about 30 min. He left the workplace with a sore throat , stuffy nose , and red eyes. At 2000, he was beginning to chill and shiver. At midnight, his temperature was 38.8°C and at 0600 the next mornin g it was 37.6°C. In the physical examin ation made at 0800 , there was nose congesti on, bilateral co njunctivit is, edema of the eye lids, and sparse crepitating rales on the left lung base. During the ensuing 10 d, the symptoms and signs gradually subsided.
Expos ure. According to the PU conveyer-belt producer , the PU was supposedly derived from toluene diiso cyanate (TDI ). In January 1988, to confirm the alleged exposure , the patient (wearing a coal-filter respirator y mask) experi mentally repeat ed, in his workshop, the imprinting job he had done in the bakery. The smoke contained 170 (0.2 m above the belt) and 15 (resp iratory zone ) ug 4,4'-MDI . rn" , but no TD!. (For the method see reference 5.) An 4,4'-MDIrelated amine, 4,4'-diaminodiphenylmethane (MDA), but not the TDI-related amine, tolu enediamine, was dete cted in hydro lyzed serum (5.6 ng · ml-I ) and urine ( 1.6 ug . g creatinine I) on 20 Novemb er 1986 [as determined by gas chromatog raphy and selected ion monitoring (unpublished observa tions and reference 6)]. Decreasing levels of MD A were found in serum and urine durin g the following months and days, respectively. (A bystand er, without respiratory protection, exposed for 24 min dur ing the aforementioned pyrolysis experiment in the workshop, had serum levels of 0.03 before the exposure and 2. FEVI.Il was again higher, 4.0 I (both times), and the vital capacity was 4.8 and 5.0 I, respectively. Skin-prick tests with isocyanate [4,4 '-MDI , phenylisocyanate (PhI), TDI , p-tolu ene monoisocyanate (P-TMI), hexameth ylene diisocyanate (HDI)] conj ugates with human serum albumin (HSA) evoked positive responses to all of the substances (wheal at least half that induced by histamin e). The skin-pric k test with 13 common a llergens was negati ve.
Routin e laborato ry tests. The erythrocy te sedimentation rate rose from 12 mm on Nove mber 20th to 22 mm on Novembe r 25th. Electrophoresis samples of serial blood reve aled acute inflammatory activity and, especially, an increase in orosomu coid . On November 20th, the neut roph ilic cell count was 7.9 . 10 9 • I-I and had, on November 2 1st, decreased to 3.3 . 10 9 • I-I. Eosinoph ils reached a maximum of 0.8· 10 9 • I-I on November 21st and a base level of 0.2· 10 9 • I-Ion Decemb er 8th. The total immunoglobulin (lg) A and total IgM were maximal on November 25th, 2.55 and 1.5 g . 1-1 , respectively, and 2.04 and 1.2 g . 1-1 , respectively, on Decemb er 12th. The total IgG (figure I) and, especiall y, the total IgE (fig ure 2) were elevated. There were no specific IgE antibodies against comm on allergens and no rheumatoid (Waa ler-Rose) or antinuclear (ANA) antibodie s. Much later, in 1993, Sjogren's syndrome (SS-A) autoantibodies were found .
The RAST values for specifi c IgE against 4,4' -MDI-H SA decrea sed after the end of the exposure (initial half-time one to two months), but they were    No precipitating antibodies against 4,4'-MDI-HSA, TDI-HSA, or HDI-HSA were found in Ouchterlony analyses (3% weight/ volume PEG6000 added to the gel). The complement proteins Clq, CIs, C3, and C4 in sera from 20 and 2 1 November 1986 were slightly increased. (For the method see reference 10). Circulating immune complexes were present above the normal range (Clq binding assay from reference II ; solid phase C l q binding assay, modified from reference 12). In vitro, the addition of 4,4'-MDI-HSA (figure 3), or TDI -HSA (not shown), in increasing concentrations up to 5 g . I-I, gave a marked increase in the Clq binding activity in the serum. With samples obtained two and four years later , no such increase was seen, but the levels of circulating immune complexes in serum were high ( figure 3).
The assessment of CI activation [crossed immunoelectrophoresis ( 13)] showed an increased concentrat ion of comple xes containing C I inhibitor, C Ir, and CIs at the time of the exposure. In the sera obtained two and four years later, increa sed levels of complexes cont aining inhibitors of both (C Ir-C Is)2 and C I were observed. Incubation of sera with 4,4'-MDI-HSA did not result in CI activation, neither were C3dg fragments generat ed. (For the method see reference 14). Furthermore, there was no cleavage of C3 after incubati on with TDI-HSA [crossed immunoelectrophoresis (15)]. Anti-Clq antibodies were found in serial serum samples collected since the exposur e, the highest concentration occu rring four years after the exposure , 280 arbitrary units· ml" (upper reference limit 16; method from reference 16).
To elucidate any constitutional deviation, we determined peripheral blood lymphocyte markers for CD2, CD3, CD4, CD8, and CD 19 using flow cytometric analysis and found them to be normal in 1990, as was the stimulation of peripheral blood mononuclear cells (PBMC) with phythemagglutinin or pokeweed mitogen. To stud y possible immunologic memor y, PBMC and B-cell depleted PBMC (Dynabeads Pan-B [CDI9], Dynal AS, Oslo) were incubated with 4,4'-MDI-HSA or with epoxy Mw 340 (DGEBA), but we found no stimulation of cells, as measured by the uptake of 3H-thymid in.
Course of events. After November 1986, the patient began to use respiratory protective devices whenever he thought he might be substanti ally expo sed to isocyanates. In January 1988 (day 428 in figur es 1 and 2), he was reexamined in the hospital because of skin rash and papular eruptions on his hand s (present for one and a half months) and facial edema (present about 10 d). In connection with the onset of symptoms, he had started to use a new glue Scand J Work Environ Health 1994, vol 20, no 3 ("super-epoxy"). In the examination, he showed an intense patch-test reaction against epoxy Mw 340, but not versus MDA, MOl , or other materials that he had used.
In March 1992, he had repeatedly heated PU materials without respiratory protection (before day 1940, figures I and 2). In April (day 1969), he imprinted on a belt (of unknown composition) in a bakery and was struck with acute respiratory distress, sore throat, chills and fatigue , lasting for about 24 h.

Discussion
Already before the first examination for symptoms (November 1986), the patient had probably had repeated isocyanate-associated systemic reactions. The high values of specific immunoglobulins already in the very first blood samples at least indicated earlier exposure(s).
A noteworthy aspect is the inadequate information supplied by the manufacturer with regard to the composition of the conveyer belt. Thus the relevant isocyanate became fully clear only after an analysis of air, serum, and urine samples. The 4,4'-MDI-related amine , MDA , was detected in hydrolyzed serum and urine . In workers exposed to MDA as such, much higher levels (in urine) have been reported (17).
Furthermore, judged from the rough exposure measurements made later in the workshop, the exposure to MDI might have been below the permissible limit (in Sweden 50 ug . rrr", time-weighted average). However, even such exposure, at least to TOl , may cause disease (2,4). In the present case, though , the exposure has been complex.
In spite of the known exposure to 4,4'-MDI only , the patient had specific IgG and IgE antibodies to a variety of isocyanates (the highest values, though, for 4,4'-MOl) and PA. This finding could be due to other exposures or to cross-reactivity, possibly because of new antigenic determinants being formed (2,(18)(19).
The specific IgE and IgG antibodies gradually disappeared after the supposed end of peak exposure to MOl, in accordance with earlier studies of IgE and IgG isocyanate antibodies (19)(20) and IgE antibodies to other small organic molecules (21). Apparently, the IgE level decreased faster than that of IgG. IgE versus TOl-HSA decayed more rapidly than IgE versus 4,4'-MOl-HSA, perhaps due to a lower affinity of the specific TDI antibodies. The decrease of the total IgE and IgG level was faster than that of the specific antibodies . Thus, in addition, there seems to have been a polyclonal activation of B cells.
We do not know with certainty that the demonstrated specific antibodies played a pathogenetic role for the respiratory and systemic reactions. IgE antibodies have earlier mainly been associated with isocyanate asthma (2)(3)(4)22), recently also with isocyanate-induced hemorrhagic pneumonitis (23). Our 220 patient had a complicated clinical picture; possibly his rhinoconjunctivitis and acute dyspnea were IgEassociated.
The patient had circulating immune complexes that bound C I q, and some increase in the complexes was observed after isocyanate conjugates were added to acute serum in vitro. This phenomenon might reflect the presence of antibodies related to the development of symptoms. The patient's sera showed the presence of excess (Clr-Cls)2 complexes, also an earlier finding for patients with chronic urticaria or angioedema (30). Furthermore, IgG binding to collagen-like fragments of Clq was demonstrated in the sera . IgG reaction with Clq has been reported for patients with systemic lupus erythematosus (16,31) and patients with hypocomplementemic urticarial vasculitis syndrome (16,32). Our patient had neither of these diseases. The pathogenetic significance of Clq antibodies is not known (32). Possibly , our findings reflect an autoimmune response to antigenic stimulation from isocyanates. Anti-C 1q antibodies have been suggested to be responsible for the majority of the solid-phase C Iq-binding IgG in the sera of most patients with systemic lupus erythematosus (31), and such antibodies in our patient probably account for part of the circulating immune complexes (32) . Decreased clearance of circulating immune complexes in the patient could not be ascribed to a deficiency of C4A, the C4 isotype most efficient in processing immune complexes (33).
This case shows that isocyanate-associated disease has to be considered in cases with unclear respiratory and general symptoms and signs. The presence of specific antibodies was the main finding, and the complement system was probably also involved . The finding of antibodies to Clq suggests an autoimmune reaction ; this reactivity may have influenced the patient' s symptoms. The presence (in 1993) of SS-A antibodies gives some further support to the possibility of an autoimmune disposition in this patient, as does his history of optic neuritis, a disease in which autoimmune mechanisms might be involved (34).
The mechanism behind isocyanate-related illness is still obscure . The present case might have a general bearing on mechanisms. It is possible that the features of our patient's disease that seem unique are not really so -but have simply not been studied earlier. Our findings should form a basis for epidemiologic approaches. More basic information concern-ing immunologic c ha racte ris tics and predisp o sin g ho st factors is c er ta inly needed.

Acknowledgments
This work w as su pp or te d b y gr ants from th e Sw edish W ork En vironment Fund, the S wedi sh Medical R e search Council (projects no 10 381 and 7921), the M edical Faculty of the University of Lund, the S wedish Nation al Associatio n ag ai ns t Rh eumatism, the Kin g Gustaf V 80th B irthday Fund, th e M Bergv all Trust, the Greta and Johan Kock Trusts , th e Profe ssor Nanna Sv artz ' Trust , th e Thelma Zoega Trust , and th e A 0sterl u nd Trust.
We thank Dr M Arborelius, Jr , MD , for hi s kind cooperation, and Mr M Adamsson for hi s skillful ass is tanc e .

Addition to proof
In 1994, th e p atient de vel oped crescentic g lo merulon ephritis w ith autoantibodie s ag ainst myel operoxidase, This devel opment un der scores his auto im m une di sposition. The s ig n if icance of hi s isocyanate re acti v it y ( in 1994 he still ha s IgE a nd IgG isocyanate antibodie s), if any , is unknown.