Document O1QVNpVdaYLpyLZJK998OMNke

Original Contributions Hepatic Disease Among Workers at a Vinyl Chloride Polymerization Plant Henry Falk, MD; John L. Creech, Jr., MD; Clark W. Heath, Jr., MD; Maurice N, Johnson, MD; Marcus M. Key, MD Eleven cases of hepatic disease, including seven cases of hepatic ang iosarcoma, have been identified to date among men employed at one vinyl chloride polymerization plant. The earliest diagnosis was made in April 1964. The two most recent cases, both angiosarcoma, were diagnosed in February 1974 as a result of systematic medical screening for liver abnor malities among workers at the plant. Ages at diagnosis have ranged from 36 to 58 years for the seven patients with angiosarcoma and from 28 to 56 years for the four patients with nonmalignant disease; durations of employment before diagnosis have ranged from 12 to 28 years and from 5 to 29 years. All 11 persons had worked in close and continuous contact with various phases f the vinyl chloride polymerization process. Review of pathologic material suggests the presence in both tumor and nontumor cases of portal fibrosis and atypical sinusoidal lining cells. A direct causal relationship between ex posure to vinyl chloride monomer and pathologic findings is postulated. (JAMA 230:59-63, 1974) CREECH and Johnson' recently re ported the occurrence of three cases of angiosarcoma of the liver among workers at a polyvinyl chloride (PVC) production plant in Louisville. Be cause this tumor is extraordinarily rare (only about 25 cases are esti mated to occur each year in the entire United States), the existence of such cases in this particular setting See also p 64. strongly suggests a causal relation ship to some phase of the PVC pro duction process. Recent animal stud ies being conducted in Italy support the concept that exposure to vinyl chloride monomer (VCM) may be the mechanism involved (C. Maltoni et al, unpublished data). Beginning in late January 1974, in- From the Cancer and Birth Detects Division, Bureau of Epidemiology, Center (or Disease Control, Atlanta, Ga (Drs. Falk and Heath), the B. F. Goodrich Company, Louisville, Ky (Dr. Creech) and Akron, Ohio (Dr. Johnson), and the National Institute for Occupational Safety and Health, Center for Disease Control, Rockville, Md (Dr. Key). Reprint requests to Cancer and Birth Defects Division, Bureau of Epidemiology, Center for Disease Control, 1600 Clifton Rd, Atlanta, GA 30333 (Dr. Falk). tensive efforts have been devoted to clinical and epidemiologic studies of past and present workers at the Louisville plant, as well as at other PVC production plants elsewhere, in order to define more precisely the ex tent of health risks among vinyl chlo ride workers. This report summarizes findings to date with respect to both malignant and nonmalignant hepatic disease among workers at the Louis ville plant, particular emphasis being given to epidemiologic features. BACKGROUND The production of PVC by polymer ization of VCM began in Germany about 40 years ago, with production in the United States starting about five years later. The industry grew rapidly after World War II, and growth has continued in recent years at a rate of about 14% per year. Cur rently in the United States, 14 plants employing about 1,500 workers pro duce VCM, while 37 plants employing about 5,000 workers polymerize PVC from VCM. Current annual produc tion of PVC in the United States is estimated at approximately 4.4 billion pounds, or 25% of world production. The B. F. Goodrich PVC polymer ization plant in Louisville first began operations in 1942. The number of persons employed at the plant di rectly in PVC polymerization has steadily increased, reaching a fairly stable level of 250 to 300 workers by the late 1950s. At present, in addition to 271 persons engaged directly in PVC polymerization, about 850 per sons are employed at the plant in other activities such as synthetic rub ber production, compounding and milling operations, managerial and clerical positions, and maintenance work outside PVC polymerization areas. Until 1966, VCM as well as PVC was produced at the Louisville plant. Since that time, however, all VCM used at the plant has been shipped by tank car from other facilities. The VCM is unloaded, stored, and then piped into large polymerization re actor vats through an essentially closed system. Each reactor receives a measured amount of VCM as well as appropriate catalysts, stabilizers, emulsifiers, and additives (and other monomeric compounds if copolymers or terpolymers are being produced), and the reaction is carried to the de sired end point. Polymerized material is dropped into secondary tanks from which unreacted VCM is recovered and recycled through a closed system; it then enters tertiary tanks from which it is concentrated, dried, and packaged. The end product consists of three different materials: (1) PVC resin (a powder with the texture of refined sugar), (2) PVC paste (a very fine powder with the texture of pro cessed flour), and (3) PVC latex (a stable suspension of PVC in liquid). For workers in PVC polymeriza tion, the point of greatest probable exposure to VCM occurs after poly- JAMA. Oct 7, 1974 Vol 230, No 1 Hepatic Disease--Falk et al 59 37.0 5^ merization, when reactors are opened and cleaned. Although the air within reactors is replaced several times be fore opening, a short burst of VCM may be released from reactors imme diately on opening. In addition, some PVC remains encrusted within react ors which, because of its porous struc ture, may retain significant amounts of entrapped VCM. In the process of chipping and cleaning this material from within reactors, retained VCM is released. Until the late 1960s, this cleaning process was done manually by a man lowered into the reactor for that purpose. Since that time, highpressure water hoses have been intro duced for cleaning reactors, and man ual cleaning is done much less fre quently. This change in work practice was designed primarily to prevent acro-osteolysis, a disease peculiar to PVC workers and characterized by Reynaud phenomenon, scleroderma like changes of the hands, and lytic lesions in the distal phalanges.2'3 CLINICAL AND PATHOLOGIC FINDINGS Table 1 summarizes salient fea tures for each case of hepatic angio sarcoma and nonmalignant liver dis ease diagnosed to date. Review of death records for plant employees dis closed a total of five cases of angio sarcoma of the liver diagnosed in PVC workers over the ten-year peri od between 1964 and 1973 (cases 1 through 5). No cases were found diag nosed before 1964, and no cases were found of hepatic carcinoma or of an giosarcoma primary at other sites. A further review of medical histories of PVC workers employed at the plant disclosed four additional men with a known history of nonmalignant liver disease demonstrated by tissue biopsy and diagnosed between 1968 and 1973 (cases 8 through 11). Because of concern aroused by the discovery of hepatic angiosarcoma in workers, a medical screening pro gram aimed at detecting hepatic ab normalities was instituted at the Louisville plant. All current employ ees, whether engaged in PVC poly merization work or not, were exam ined (see p 64). As a result of this program, two additional cases of an giosarcoma of the liver were diag Case No. Hepatic Tumors 1 Age at Diagnosis/ Race/Sex 52/W/M Month and Year of f i i . .i Diagnosis Death 4/64 4/64 2 43/W/M 8/67 1/68 3 36/W/M S/70 9/71 4 49/W/M 3/73 3/73 5 58/W/M 12/73 12/73 Table 1.--Clinical Initial Symptoms 8/63-Fatigue 1/64-Right upper quadrant pain 9/66-Fatlgue, "pleurisy" 8/67-Epigastrlc "knot" and pain 1/70 and 5/70-Melena 12/63 and 5/65-Melena and hematemesls 7/73-Weakness and weight loss 6 45/W/M 7 43/W/M Nonmalignant Hepatic Disease 8 46/W/M 9 28/W/M 10 56/W/M 11 S6/W/M 2/74 2/74 12/68 1/72 9/73 9/73 8/73-Fatlgue, "pleurisy" Asymptomatic 10/68 and 11/68-Melena 9/71-Chest pain, weight los^^ 2/73-Hospitalized for hernilWF repair, icterus noted 9/73-Hospitalized for chole lithiasis nosed, both in PVC workers (cases 6 and 7). At the present time, no cases of angiosarcoma of the liver or of nonalcoholic liver disease as demon strated by biopsy have yet been iden tified among non-PVC employees at the plant. All patients have been white men. Cases of Hepatic Angiosarcoma Ages at diagnosis for the seven pa tients with tumors (cases 1 through 7) have ranged from 36 to 58 years (average age, 46.7). Initial clinical features have varied widely; there were no signs or symptoms in case 7. Four patients (cases 1, 2, 5, and 6) were first seen with weakness and tiredness; two of these also had inter mittent pleuritic pain. These symp toms by themselves were not suffi ciently severe to warrant medical evaluation until the appearance of acute abdominal pain, pronounced weight loss, or abnormal findings on serologic screening. Two patients (cases 3 and 4) were entirely asymp tomatic until the abrupt onset of gas trointestinal bleeding. Two patients had clinically obvious hepatosplenomegaly (cases 2 and 5), but four (cases 3,4,6, and 7) had normal physi cal findings. While all seven patients had evidence of liver function abnor mality at time of initial work-up, no consistent pattern emerged, and in several instances abnormalities were only slight (cases 2, 3, 4, 6, and 7). In several cases, relatively mild hepatic dysfunction coexisted with either faradvanced portal hypertension (cases 3 and 4), unresectable angiosarcoma 60 JAMA, Oct 7, 1974 Vol 230, No 1 SL 028321 Hepatic Disease--Falk et al and Pathologic Findings of Cases of Liver Disease Physical Findings 1 /64-Right upper quadrant tenderness 8/67-Epigastric mass and splenomegaly 1/70-No abnormalities 5/70-Hepatosplenomegaly 12/63-No abnormalities 5/65-Splenomegaly 7/73-Hepatosplenomegaly 2/74-No abnormalities 2/74-No abnormalities Hepatic Work-up* 1/64-Moderate elevation in TB. SGOT 8/67-Mild elevation in AP, SGOT, LDH Platelet count, 33.000/cu mm Liver scan: large defect, splenomegaly 5/70-Mild elevation in TB, AP, SGOT, LDH Liver scan: large defect Esophogram; varices 5/65-Elevated SGOT Esophagoscopy: varices 7/73-Marked elevation in AP Mild elevation in TB, SGOT Liver scan: diffuse disease, hepatosplenomegaly 2/74-Mild elevation In LDH Liver scan: possible defect 11/73-Mild elevation in TB, AP Liver scan and angiogram: 4-cm defect Pathologic Findingst 1/64-OLB: slight focal hepatitis 3/64-NLB: unchanged 4/64-PM: hepatic angiosarcoma with metastases 8/67-OLB: angiosarcoma 1/68-PM: hepatic angiosarcoma with spread to diaphragm and abdominal wall 5/70-OLB: angiosarcoma No PM 5/70-OLB; toxic hepatitis 10/70-NLB: hepatitis, cirrhosis 10/70-OLB: slight hepatitis 3/73-PM: hepatic angiosarcoma 7/73-NLB: fibrosis OLB: periportal inflammation and fibrosis 12/73-PM: hepatic angiosarcoma spread to duodenum 2/74-OLB: angiosarcoma 2/74-OLB: angiosarcoma and extensive portal fibrosis, subcapsular fibrosis 11 /68-Splenomegaly 9/71-Hepatosplenomegaly ^B/73-Splenomegaly 9/73-No abnormalities 10/68-Marked elevation in AP, BSP Mild elevation in SGOT Esophagram: varices 9/71-Mild elevation In TB, SGOT 10/71-Liver scan: splenomegaly 3/73-Moderate elevation in TB. AP Liver scan: splenomegaly, small liver 9/73-Moderate elevation in SGOT Mild elevation in TB, LDH 11/68-OLB: portal fibrosis, subcapsular fibrosis 12/71-OLB: portal fibrosis, subcapsular fibrosis 9/73-OLB: slight portal fibrosis 9/73-OLB: chronic hepatitis with focal fibrosis *TB indicates total bilirubin; SGOT, serum glutamic oxaloacetic transaminase; AP, alkaline phosphatase; LDH, lactic dehydrogenase; BSP, sulfobromophthalein. tOLB indicates open-liver biopsy; NLB, needle liver biopsy', PM, postmortem examination. (case 6), or angiosarcoma with exten sive portal fibrosis (case 7). In cases 1 and 2, large hepatic masses were present at initial evaluation. The con ditions of three patients (cases 1, 4, and 5) were not diagnosed until au topsy, despite multiple liver biopsies. Preliminary pathologic review sug gests that in all 5 patients for whom nonmalignant hepatic tissue is avail able (cases 2, 3, 4, 5, and 7), similar nonmalignant hepatic lesions exist, consisting of portal fibrosis, sinusoi dal dilation, and atypical sinusoidal lining cells. Conceivably, such lesions may represent a precursor stage in the development of hepatic angiosar coma. Nonmalignant Hepatic Disease Ages at diagnosis for these four pa tients (cases 8 through 11) range from 28 to 56 years (average age, 46.5). Ini tial clinical manifestations varied widely: one patient (case 8) had gas trointestinal bleeding; one (case 9) had chest pain and weight loss; and two patients (cases 10 and 11) had unrelated problems. On physical ex amination, hepatosplenomegaly was present in one patient (case 9), sple nomegaly alone in two (cases 8 and 10), with normal findings in one pa tient (case 11). Three patients under went splenectomy either for marked splenomegaly (case 9) or as part of splenorenal shunt procedures (cases 8 and 10). Results of liver function tests varied widely and showed no consis tent patterns in relation to clinical manifestations. All four patients were found on liver biopsy to have some degree of hepatic fibrosis. Pathologic review of specimens available from these sug gests a close histologic similarity to the manifestations of portal fibrosis and sinusoidal changes described in the cases of angiosarcoma. Three pa tients (cases 8, 9, and 11), as well as one patient in the angiosarcoma group (case 7), were found at surgery to have a peculiar white, speckled ap pearance to the surface of the liver, which on pathologic section was seen to reflect diffuse subcapsular fibrosis, EPIDEMIOLOGIC FINDINGS The seven men with angiosarcoma had been employed at the plant for 12 to 28 years (average, 18.0), and the four with nonmalignant disease, be tween 5 and 29 years (average, 20.6) JAMA, Oct 7, 1974 Vol 230, No 1 SL 028322 Hepatic Disease--Falk et al 61 Table 2.--Duration of Employment for Patients With Angiosarcoma of Liver and Nonmalignant Hepatic Disease Casa No. Duration of Employment, PVC-Poiymerization Buildings Other Buildings* yr/mo A BCD E F G H 1 J K L Others 1 19/8 8/11 0/3 0 0 8/5 2/1 00 0 0 00 0 2 17/11 0/1 12/5 2/0 0 0/6 0 2/9 0/2 0 0 0 0 0 3 13/1 1/2 1/7 5/5 4/4 0 0/3 0 0 0/2 0/200 0 4 16/5 1/2 14/9 0/6 0 0 0 0 0 0 0 0 0 0 5 27/7 0/6 2/8 0 24/0 0/5 0 0 0 0 0 0 0 0 6 12/0 0 0/1 2/5 9/6 0 0 0 0 0 0 0 0 0 7 19/0 2 wk 2 wk 12/1 5/5 0 1/4 0 0 0 0/2 0 0 0 8 24/5 4/1 4/1 0 14/0 1/4 0 0 0 0 0 0/11 0 0 9 5/6 i wk 1 wk 3/4 2/0 0 0 0 0 0/1 0 0 0 0 10 23/6 0 19/6 1/4 0 1/5 0 0 01/2 0 0 0/1 0 11 <1 Total 28/11 208/0 21/8 7/1 0/2 0 0 0 0 0 0 0 0 0 37/8 62/6 27/3 59/3 12/1 3/8 2/9 0/2 1/5 0/4 0/11 0/1 0 0 No. of yr in operation 32 32 30 27 26 27 27 29 19 32 26 32 26 32t Total No. of employees in 1973 36 36 48 40 14 1 12 0 52 72 18 8 ~600 *ln Building E, PVC drying and packaging are done; in F, alcohol synthesis (formerly monomer synthesis); G. PVC chlorination; H, no longer in operation (formerly monomer synthesis); 1, compounding and milling; J, synthetic rubber; K, compounding (formerly warehouse and receiv ing); L, PVC drying. tOn the whole, other buildings were in operation as long as the plant as a whole. Table 3--Work Histories for Patients With Angiosarcoma of Liver and Nonmalignant Hepatic Disease Case No. 1 2 3 4 5 6 7 8 9 10 11 First Liver Disease Diagnosis, Date April.1964 Aug 1967 May 1970 March 1973 Dec 1973 Feb 1974 Feb 1974 Dec 1968 Jan 1972 Sept 1973 Sept 1973 Total Duration of Work Before Diagnosis, yr/mo 19/8 17/11 13/1 16/5 27/7 12/0 19/0 24/5 5/6 23/6 28/11 Total Duration of Work as Chemical Halper In PVC-Polymeri zation Buildings, mo 6 2 65 20 8 43 44 47 65 0 9 (Table 2). Ten of the 11 patients (cases 2 through 11) worked exclu sively or predominantly in one or more of the four PVC polymerization buildings at the plant (Table 2, build ings A, B, C, and D). Patient 1 worked about half of his total employ ment time in a PVC polymerization building (building A) and for an al most equal time in a separate PVC drying and packaging building (building E). Three patients never worked elsewhere than in the PVC polymerization buildings (case 4, 6, and 11). As can be seen from Table 2, ization (buildings E through L and all others) were only sparsely repre sented in relation to total employ ment. For the four polymerization buildings that in 1973 employed a to tal of 160 persons, 187 man-years of employment are represented among the 11 cases. For the remainder of the plant with a total efaiployment of about 950 persons in 1973, only 21 man-years were recorded. Buildings A and B, opened in 1942 and 1944, respectively, are the older of the four polymerization buildings building C (1947) and building E C and D have more reactors (48 each) than either building A (35 reactors) or building B (25 reactors); the newer reactors have approximately a onethird greater capacity. Employment among all 11 patients involved all four polymerization buildings. Three patients had worked in all four buildings; five had never worked in building D, three had never worked in building A or building C, and although all 11 had worked at some time or another in building B, three had worked there less than one month. It appears unlikely that some chemical or procedure unique to any one building can be implicated as a causative factor; almost twice as many man-years of employment, however, were represented among cases for buildings B and D (121 years, 9 months) as for buildings A and C (64 years, 11 months). While this difference may or may not be meaningful, some variations do exist between the various buildings that could conceivably be important as risk factors. Buildings A and C pro duce homopolymer resin exclusively while various copolymers and terpolymers are produced in building B, and all PVC paste and almost all PVC la tex is produced in building D. Those workers who are probably 62 JAMA. Oct 7, 1974 Vol 230. No 1 028323 Hepatic Disease--Falk et al helpers whose principal job is to clean reactors. As can be seen in Table 3, ten of the 11 patients worked at some time as helpers. While virtually every employee at the plant has worked as a helper before being promoted to more advanced work, the average work du ration (ie, time from starting work at the plant to date of diagnosis) was 23.5 years for the five patients who spent nine months or less as helpers and 15.1 years for the six patients who spent 20 months or more. This suggests indirectly a possible rela tionship between intensity of expo sure and latent period for liver dis ease. All 11 patients, or members of their immediate families, were indi vidually interviewed regarding past hepatic disease and possible exposure to hepatotoxic agents. None of the patients had a history of hepatitis or of exposure to hepatitis, and none had taken hepatotoxic drugs. Three pa tients (cases 1, 7, and 9) may have had significant alcohol intake. None, except for the patient in case 7, re called exposure to possible hepato toxic chemicals outside the work en vironment, in particular to either arsenic or thorium dioxide, two chem icals previously implicated as causes of hepatic disease and hepatic angio sarcoma in humans.4-7 Patient 7 gave a history of exposure to arsenical in secticides on the family farm be tween the ages of 6 and 15 years; he both mixed and sprayed the insecti cides two to three times a year for about three hours on each occasion. In no case was there any history of acroosteolysis. COMMENT Before the report by Creech and Johnson,1 the only evidence that VCM might be oncogenic came from ani mal experiments. In 1971, Viola et al8 published data suggesting oncogenic ity of VCM when inhaled by rats at very high doses; tumors of many tis sue sites, including lung, bone, and skin were recorded. Preliminary re sults of a more recent animal study in Italy by Maltoni et al (unpublished data) appear to indicate that angio sarcoma of liver as well as of other tissues can be induced in rats by at mospheric levels of VCM that are not uncommon in the human workplace environment. In light of these obser vations, it appears likely that expo sure to VCM is responsible for the Louisville cases. Further support for this hypothesis is needed, of course, from additional epidemiologic data concerning workers at other PVC and VCM plants. Further toxicologic analyses are also needed to address the possibility that the active onco genic material may be some metabo lite of VCM instead of VCM itself. Whatever the precise mechanism for oncogenicity and hepatotoxicity of VCM, the Louisville data suggest that relatively high levels of VCM ex posure and relatively long intervals since first exposure (20 years or so) are involved. This does not rule out the possibility of less marked health effects at lower doses or at shorter in tervals of exposure, but it does for the present focus attention on the imme diate problem of assessing the health status of persons exposed in the re mote past to high doses. None of the Louisville cases involved men work ing at the plant less than six years before diagnosis, and it can be safely assumed that levels of VCM exposure during earlier years of PVC produc tion were considerably higher than at present because of less stringent work practice procedures and less at tention to minimizing possibilities of VCM exposure in places of work. In humans, both thorium dioxide and arsenic have previously been re ported as causes both of hepatic dis ease and of angiosarcoma of the liver. In only one case at the Louisville plant (case 7) was there any history of exposure to either of these two ma terials (arsenic in insecticide spray). Likewise, there is little or no evidence among the Louisville cases that ex cessive alcohol intake plays any accel erating or potentially cocarcinogenic role or that any direct relationship exists between acro-osteolysis and liver disease. Various data now suggest that VCM (or some derived metabolite) may produce in addition to angiosar coma of the liver a nonmalignant he patic disorder characterized by portal fibrosis and portal hypertension. This is suggested by the fact that such fi brosis was present in at least five of the Louisville tumor cases and was in addition observed in four other Louis ville vinyl chloride workers without tumor. Recent observations in Ger many," together with earlier reports from Eastern Europe,suggest that hepatic fibrosis and portal hyperten sion represent an occupational dis ease not uncommon among vinyl chlo ride workers. Conceivably, such fibrotic liver disease represents a premalignant state. If this proves to be so, the early detection of such liver disease may be of greater industrial and public health importance than de tection of tumor itself, both because hepatic fibrosis may well emerge as a more frequent condition than tumor in vinyl chloride workers, and because it remains a possibility that very early liver abnormalities may be re versible or nonprogressive after workers have been removed from high-risk areas. Laszlo Makk, MD, Lynn Ogden, MD, Edward Fadell, MD, Anne Riehman, MD, Curtis Song ster, MD, George Sehrodt, MD, Edward Callahan, MD, Jerry Clanton, MD, James Kurfees, MD, Stanley Seipel, MD, and Will Ward, MD, sup plied records and pathologic materials. Louis Thomas, MD, and Hans Popper, MD, re viewed pathologic material. References 1. Creech JL Jr, Johnson MN: Angiosarcoma of liver in the manufacture of polyvinyl chloride. J Occup Med 16:150-151, 1974. 2. Wilson RH, McCormick WE, Tatum CF, et al: Occupational acro-osteolysis. JAMA 201:577578,1967. 3. Dodson VN, Dinman BD, Whitehouse WM, et al: Occupational acro-osteolysis: III. A clinical study. Arch Environ Health 22:83-91, 1971. 4. Da Silva Horta, Cayolla da Motta L, Abbatt JD, et al: Malignancy and other late effects fol lowing administration of Thorotrast. Lancet 2201-205, 1965. 5. Da Silva Horta, Cayolla da Motta L: Fol lowup study of thorium dioxide patients in Por tugal. Ann NY Acad Sci 145*30-842, 1967. 6. Regelson W, Kim U, Ospina J, et al: Hemangioendothelial sarcoma of liver from chronic arsenic intoxication by Fowler's solution. Cancer 21514-522, 1968. 7. Morris JS, Schmid M, Newman S, et al: Ar senic and noncirrhotic portal hypertension. Gas.troenterology 64:86-94, 1974. 8. Viola PL, Bigotti A, Caputo A: Oncogenic response of rat skin, lungs, and bones to vinyl chloride. Cancer Res 31:516-522, 1971. 9. Marsteller HJ, Lelback WK, Muller R, et al: Chronisch-toxische Leberschaden be! Arbeitem in der PVC-Production. Dtsch Med Wochenschr 98:2311-2314, 1973, 10. Suciu F, Drejman I, Valaskai M; Etude des maladies dues au chlorure de vinyle. Med Lav 58:261-271, 1967. JAMA, Oct 7, 1974 Vol 230, No 1 St 028320 Hepatic Disease--Falk et al 63 Liver Damage and Angiosarcoma in Vinyl Chloride Workers A Systematic Detection Program Laszlo Makk, MD; John L. Creech, MD; Joseph G. Whelan, Jr., MD; Maurice N. Johnson, MD Recently, five cases of hepatic angiosarcoma were found in Louisville. All of the patients had worked in a plant engaged in the production of polyvi nyl chloride (PVC) from vinyl chloride. A systematic health surveillance pro gram was instituted to detect existing liver damage or angiosarcoma and to find early signs of developing liver damage. A protocol was devised that be gan with on-site phlebotomies done in the plant and progressed to sophis ticated hospital studies, if important abnormalities were found. All of the 1,183 employees of the plant were tested. Two new cases of he patic angiosarcoma and three cases of portal fibrosis were discovered by this program at the initial screening, it is hoped that continuous health monit ring will aid detection of liver abnormalities while still in a reversible or curable state. (JAMA 230:64-68, 1974) ANGIOSARCOMA of the liver is one of the rarest human malignant neo plasms. The annual incidence in the United States is about 21 cases per year. Recently, seven cases of angio sarcoma were discovered in Louis ville. These patients were scattered in various hospitals in the area and were cared for by their personal physi cians. Initially, one of us noted the diagnosis of angiosarcoma of the liver on a death certificate and recalled that three years earlier he had per formed a biopsy of a liver with the same diagnosis. After he found that both of these patients had worked in a plant engaged in the production of polyvinyl chloride (PVC), careful data collection was begun. Thorough From the departments of pathology (Dr. Makk), surgery (Dr. Creech), and radiology (Dr. Whelan), St. Anthony Hospital, and the Medical Department (Drs. Creech and Johnson), B. F. Goodrich Co., Louisville. Reprint requests to 1313 St. Anthony Place, Louisville, KV 40204 (Dr. Makk). search of plant and area hospital medical records resulted in the dis covery of three more cases from ne cropsy reports recorded over a tenyear period. Two additional cases were diagnosed by biopsy in the past two months as a result of this pro gram. See also p 59. Each of the patients with angiosar coma who were under our care, and some without angiosarcoma, had fi brosis of portal areas and other path ologic changes apparent on liver biopsy specimens. None of these pa tients had any specific abnormal find? ings on physical examination or rou tine laboratory tests. In view of this, a systematic detection program was devised and instituted with three ob jectives: (1) to find any presently ex isting liver damage or angiosarcoma in workers exposed to vinyl chloride or its polymers, (2) to detect early liver damage, if possible while in a re versible state, with continuous health monitoring, and (3) to provide infor mation for preventive measures. Materials and Methods A program was designed whereby large numbers of workers could be screened for liver disease with rela tive ease. The significant abnormal ities were evaluated by a combined clinical, comprehensive laboratory, and roentgenographic study. The re sults of this .evaluation served as a basis for the treatment of the pa tients (Fig 1). The initial laboratory procedure for detection was a 12- or 18-factor automated chemical analy sis. The 12-factor analysis consisted of tests for the following: calcium, in organic phosphorus, glucose, blood urea nitrogen (BUN), uric acid, cho lesterol, total protein, albumin, total bilirubin, alkaline phosphatase, lactic acid dehydrogenase (LDH), and se rum glutamic oxaloacetic transami nase (SCOT). The 18-factor analysis consisted of the preceding plus cre atine phosphokinase (CPK), creati nine, and serum electrolyte determi nations. The blood was drawn at the plant's Medical Department; trans portation, centrifugation, and test determinations were done promptly. If the initial laboratory results were normal, the tests were repeated in three months. If abnormalities not related to the liver were found by any of the screening procedures, they were followed up according to their severity. If one liver-related abnor- 64 JAMA. Oct 7, 1974 Vol 230, No 1 SL 028325 Liver Damage--Makk et al Fig 1.--Liver disease detection protocol. mality was found, the 12-factor anal ysis was repeated in three weeks. If the single abnormality presisted, or if two or more liver-related abnormal ities were found on the initial labora tory tests, the patient underwent a comprehensive examination consist ing of tests for the following values: a 12-factor automated chemical analy sis, serum electrophoresis, LDH isoenzymes, fractionated alkaline phosphatase, direct and indirect bili rubin (if total level was elevated), se rum glutamic pyruvic transaminase (SGPT), y-glutamic transpeptidase (GGTP), isocitrate dehydrogenase (ICD), a-fetoprotein (fetoglobulin), and carcinoembryonic antigen (CEA). Also, a complete blood cell count (CBC), platelet count, chest roentgen ogram, and liver and spleen scan were done. Abnormalities found in the initial 12-factor chemical analyses indicat ing the need for comprehensive pro Table 1.--Liver Enzyme and Bilirubin Results Test Alkaline phosphatase, mU (milliunits)/ml y-Glutamlc transpeptldase, mU/ml Serum glutamic oxaloacetic transaminase, mU/ml Bilirubin, mg/100 ml Serum glutamic oxaloacetic transaminase, mli/ml Isocltrate dehydrogenase, mU/ml Lactic dehydrogenase. mU/ml No. Of Patients Studied 72 70 68 72 73 59 72 No. (%) With Abnormal Values 35 (49.4] 31 (44.3) 19 (29.8) Normal Range 3-85 6-28 4-25 19 (26.4) 13 (18.0) 0.15-1.0 12-40 9 (18.0) 8 (11.1) 0-7 90-225 Abnormal Range 87-135 29-575 27-559 1.1-2.6 48-150 8-88 237-475 files were grouped according to the following work areas: (1) PVC pro duction, (2) synthetic rubber produc tion, and (3) all others (including maintenance, shipping, laboratory, and other salaried employees). Liver Abnormalities Dictated Hospital Studies If abnormal scan or other findings suggested serious liver disease or the possibility of angiosarcoma, the pa- JAMA, Oct 7, 1974 Vol 230. No 1 Liver Damage--Makk et al 65 028326 St- Table 2.--Twelve-Factor Analysis Abnormalities and Comprehensive Screening Cases According To Work Area Work Area Polyvinyl chloride production Synthetic rubber production All other Total 12-Factor Analysis Total No. 274 203 706 1.183 No. [%) With ' Abnormality S9 (21.5) 58 (28.6) 189 (26.7) 306 (26.7) No. (%) Compre hensive Profile Needed 26 (9.8) 14 (6.9) 35 (4.9) 75 (6.3) # Case No. 1 2 3 4 5 6 7 Diagnosis Angiography Angiosarcoma Tumor right lob< cirrhosis Angiosarcoma Tumor, both lob. and cirrhosis Portal Fibrosis, fatty change ... Portal Fibrosis, moderate Normal Portal Fibrosis, slight Normal Normal Normal Normal *f Indicates elevated values; I, depressed values. tN Indicates normal values. Fig 2.--Technetium 99m sulfur colloid liver scan, anterior view. Typical filling defect in right lobe of liver. Fig 3.--Hepatic arteriogram, six seconds. Circumferential tumor vessels are present in right lobe of liver (arrows). Main hepatic artery is normal size. Intrahepatic vessels are small and tortuous suggesting cirrhosis. Fig 4.--Hepatic arteriogram, ten seconds, portal venous phase; Circumferential tumor stain is present with a central area of radiolucency (arrows). Stain persisted through the late portal hepatogram phase, up to 34 seconds. tients were hospitalized and hepatic panangiography was performed, con sisting of hepatosplenic arterio gram and phlebogram with free and wedged hepatic pressure measure ments. The hepatic arteriogram was performed by the Seldinger tech nique.1 A mixture of diatrizoate so- -- dium and diatrizoate meglumine (Renografin-76) was used as contrast medium. If the angiograms demon strated findings suggestive of a tu mor, open-liver biopsy was advised with consideration of hepatic lobec tomy, if so indicated at exploration. If angiosarcoma was limited to one lobe, a biopsy specimen of the unin volved lobe was obtained first to eval uate the extent of fibrosis or other damage and render diagnosis by fro zen section. Then, a biopsy specimen of the tumor was obtained and stud ied likewise. If the uninvolved lobe had no no table disease, lobectomy of the neo plastic lobe was to be performed. If the nonneoplastic lobe was diseased (ie, showed fibrosis to a degree that lobectomy might have led to liver failure), or if the angiosarcoma in volved both lobes, chemotherapy was to be considered. On biopsy, a wedge and deep Vim-Silverman needle specimen was obtained from each lobe when no gross tumor was found. In the presence of neoplasm, only wedge biopsies were taken. Results Screening profiles were obtained from 1,183 employees by the 12- or 18- 66 JAMA, Oct 7, 1974 Vol 230, No 1 SL 028327 Liver Damage--Makk et al Table 3.--Tissue, Roentgenologic, and Laboratory Examination Correlation 9 Liver Scan Filling defect and cirrhosis Alkaline Phosphate f y-Glutamlc Transpetidase tt Level of Serum Glutamic Pyruvic Transaminase T Bilirubin t Filling defect, both lobes t T NN Suggestive of cirrhosis t t t t Suggestive of cirrhosis t t t N Suggestive of cirrhosis t t NN Normal t N NN Normal t NNN Serum Glutamic Oxaloacetic Transaminase t t t N N N N Lactic Dehydrogenase Nt t N N N N N Platelet Count l t N N N N N factor chemical analysis. Of these, 75 (6.3%) had either two liver-related ab normalities on the initial screening or one such abnormality that persisted. These patients had' comprehensive evaluations. One patient had initial screening in the plant, further eval uation and biopsy elsewhere, and then was transferred to our institu tion for further studies. Results of al kaline phosphatase, GGTP, SGPT, bil irubin, SGOT, ICD, and LDH examinations are listed in Table 1, Significant Biochemical Findings Other tests of the comprehensive liver profile yielded the following sig nificant results; moderately elevated total protein values were found in three patients with a similar eleva tion in y-globulin content. Serum electrophoresis showed slight to mod erate increase in y-globulin in 11 pa tients. The /3-globulin value was slightly lowered in six patients. In 32 patients with elevated alka line phosphatase values, 15 had in creased liver fractions. In 37 patients with normal' levels of alkaline phosphatase, fractionation was also performed. Thirteen showed relative elevation of liver fractions. Eight of 72 patients (11.1%) showed elevated values for total LDH. Four of these had isoenzyme 4; two, isoenzyme 5; and one, isoenzyme 3 elevation; one had normal isoenzyme proportions. Of 64 cases with normal LDH val ues, 44 showed relative elevation in level of isoenzyme 4 (68.7%); two, rela tive isoenzyme 5; and one, relative isoenzyme 1. Fetoglobulin determinations gave normal results in all cases. Results of carcinoembryonic antigen radio immunoassays were normal in all pa tients, with one marginal result in a patient who smoked. Complete blood cell counts disclosed normal values for hemoglobin, hematocrit, red blood cell (RBC) count, and corpuscular indexes in each patient. Four patients had slight to moderate granulocytosis. One patient showed relative lym phocytosis. Twenty-three patients showed slight, and seven moderate, monocytosis. Three patients had mod erately increased eosinophil counts. One patient had thrombocytopenia with a platelet count of 98,000/cu mm, and another had thrombocytosis with a count of 750,000/cu mm (nor mal range, 140,000 to 440,000). The number of 12-factor abnormalities and comprehensive profiles according to work areas is listed in Table 2, Radiological Studies Seventy-four patients had liver scans: 11 of these scans (14.8%) were abnormal, and 2 were borderline. Six patients had filling defects (Fig 2). Three of these and another five pa tients had irregularities suggestive of cirrhosis. Spleen scans were per formed concurrently with liver scans in 71 patients. Three patients had had splenec tomies in the past. Nine patients (12.7%) had enlargement of the spleen to 22 cm (normal spleen size is consid ered less than 14 cm in length'2). Splenomegaly with abnormal liver scan was present in three patients. Hepatic panangiography was per formed in seven patients. By pan angiography, two of the liver scan filling defects were found to be tu mors (Fig 3 and 4). The anomalous po sition of the gallbladder in one case, and dilatated short gastric and collat eral portal veins in another correlated with the defects on the scans. The other angiograms were normal. Of the six splenic arteriograms, two showed splenomegaly; one of these with multiple splenic aneu rysms and the other, intrasplenic ar terial strictures. The other four were normal. Surgical Findings Exploratory laparotomy was per formed in seven patients; in two, an giosarcoma was found. Resection could not be carried out in one patient because of the extent of neoplasm. In the other patient, angiosarcoma was limited to one lobe, but because of portal fibrosis in the other lobe, resec tion was deferred. In three patients, varying degrees of portal fibrosis was found on lap arotomy. Of the remaining two pa tients, one had a cholecystectomy, the other an incisional hernia repair. Liver biopsy was performed con comitantly on each. Both of these liv ers were essentially normal on biopsy and on laboratory testing. Correla tion of biopsy, roentgenographic, and laboratory findings is given in Table 3. JAMA, Oct 7. 1974 Vol 230, No 1 SL 028328 Liver Damage--Makk et al 67 Chest roentgenograms and labora tory examinations not related to liver function disclosed only occasional ab normalities. These were followed up appropriately. All were within the findings expected from a screening of this size male adult population. No lung tumors were found. Comment Nonneoplastic liver damage in polyvinyl chloride production workers has been reported.3'1 Acro-osteolysis was also observed in such workers.5 Viola et al6 reported the development of skin, lung, and bone neoplasms in rats exposed to vinyl chloride. Maltoni7 reported hepatic angiosarcomas and other neoplasms in rats exposed to high doses of vinyl chloride. To our knowledge, the Louisville cases are the first recognized human angiosar comas of the liver associated with working in vinyl chloride monomer polymerization. Creech and Johnson8 reported the first angiosarcoma of the liver among vinyl chloride workers. Block8 also reported on the subject; one of our cases is included in his re port. The report of Falk et al (p 59) on the epidemiology of this disease also includes two cases of angiosar coma discovered with our program. As the results indicate in Table 2, the percentage of abnormalities on the 12-factor examination seems higher than would be expected among a presumably healthy working popu lation. These results also indicate that a somewhat smaller proportion of persons engaged in PVC production had abnormalities on initial 12-factor analysis than either those who worked in synthetic rubber produc tion or all other workers in the plant. However, considerably larger num bers and percentages of the PVC workers had liver disease serious enough to warrant comprehensive profiles. All the patients who needed angiography, the two patients with angiosarcoma, and two of the three with portal fibrosis had worked in PVC production. In our experience, the logistics of this program worked well from on site screening to sophisticated test ing. This surveillance discovered two new, otherwise unsuspected angiosar comas of the liver and three cases of portal fibrosis. The good correlation between normal liver biopsies and es sentially normal liver function tests in cases 6 and 7 was reassuring. In addition this program provided a wealth of base-line liver function test results and other health information. These data formed the basis of a con tinuing health monitoring program for these workers. GGTP Reflects Extent of Damage Of the liver function tests, the GGTP determination seemed to be most useful in detecting abnormal ities and reflecting the extent of liver lesions by the degree of elevation. Alkaline phosphatase, SGPT, SGOT, LDH, and bilirubin tests were also useful. The large number of LDH1 isoenzyme elevations are noteworthy, but they merit further study before specific significance can be attached to them. Fetoglobulin and carcinoembryonic antigen determinations were normal. No test was specific for angiosarcoma. The liver function studies appeared to reflect quite well the extent of liver damage, except in patient 2, where this correlation was poor. He had extensive angiosarcoma and only slight elevation in LDH val ues on the initial screening that per sisted. This prompted his astute per sonal physician to request more sophisticated studies and biopsy ex amination. It is therefore quite im portant than even borderline abnor malities be carefully followed up. Scans and Angiograms Liver scans and angiographic stud ies corresponded well with results of laboratory tests and biopsy examina tions in the detection of cirrhosis, even in early stages. When roentgenographic studies indicated cirrhosis, the concomitant histopathologic ex amination disclosed a peculiar fibro sis of the portal areas instead of cir rhosis. It is our understanding that when communications refer to cir rhosis in the context of this condition, the histological findings are those of a peculiar portal fibrosis and Kupffer cell hyperplasia. Hepatic panangi ography with pressure measurements as described by Viamonte et al' was very helpful in the differentiation of neoplastic from nonneoplastic lesions that appeared as a filling defect on scanning. It also proved helpful in the preoperative evaluation of the extent of the tumor. In addition to angiogra phy in selected cases, we are planning liver biopsies through the right jugu lar vein as described by Rosch et al.11 The experience gained from this study enabled us to modify our method. We added GGTP and SGPT determinations to the 12-factor initial examination. Tests for fetoglobulin, ICD, and carcinoembryonic antigen as well as electrophoresis were de leted from our comprehensive profile. Until specific tests for hepatic angio sarcoma or vinyl chloride-induced liver damage are available, in our ex perience such a surveillance program of laboratory testing, roentgenographic examinations, and clinical evaluation will help to detect liver damage in an early, possibly rever sible or curable state. James T. Kurfees, MO, Anchorage, Ky, gave permission for inclusion of the results from one of his patients in this study. References 1. Seldinger SI: Catheter replacement of the needle in percutaneous arteriography. Acta Ra diol 39:368-376, 1953. 2. Siegel RM, Becker DV, Hurley JR: Eval uation of spleen size during routine liver imag ing with **"Tc and scintillation camera. J Nucl Med 11:689-692, 1970. 3. Marsteller HJ, et al: Chronisch-toxische-leberschaden bei arbeitem in der PVC-production Dtsch Med Wochenschr 98:2311-2314, 1973. 4. Suciu T, Drejam T, Valaskai M: Etude des maladies dues au chloride de vinyl. Med Lav 58:261-271, 1967. 5. Wilson RH, et al: Occupational acroosteolysis: Report of 31 cases. JAMA 201:577, 1967. 6. Viola PL, Bigotti A, Caputo A: Oncogenic response of rat skin, lung, and bones to vinyl chloride. Cancer Ret 31:516-522, 1971. 7. Maltoni C: Exhibit 6: Judge Myatt's hearing on oncogenic potential of vinyl chloride 43-63, 1974. 8. Creech JL, Johnson MN: Angiosarcoma of liver-in the manufacture of polyvinyl chloride. J Occup Med 16:150-151, 1974. 9. .Block JB: Angiosarcoma of the liver follow ing vinyl chloride exposure. JAMA 22953-54, 1974. 10. Viamonte MJ, Warren WD, Foreman JJ; Liver panangiography in the assessment of por tal hypertension in liver cirrhosis. Radiol Clin North Am 8:147-167, 1970. 11. Rosch J, et al: Transjugular approach to liver biopsy and transhepatic cholangiography. N Engl J Med 289:227-231. 1973. 68 JAMA, Oct 7, 1974 Vol 230, No 1 SL 028329 Liver Damage--Makk et al