347
ASPIRIN-INDUCED HEPATOTOXICITY IN JUVENILE RHEUMATOID ARTHRITIS A PROSPECTIVE STUDY BALU H. ATHREYA, GLENDA MOSER, HENRY S. CECIL, and ALLEN R. MYERS
Recent reports of and our own experience with biochemical alterations of liver function secondary to salicylate therapy stimulated this prospective study. Thirty-four children with juvenile rheumatoid arthritis, 6 children with acute cartilagenous necrosis of the hip following slipped capital femoral epiphysis, and 2 children with ulcerative colitis and hip disease who were on salicylates were followed over a period of 1-27 months with serial determinations of serum glutamic oxaloacetic transaminase (SGOT), serum glutamic pyruvic transaminase (SGPT), lactic dehydrogenase (LDH), alkaline phosphatase (AP), bilirubin, and serum salicylate. Prothrombin time was measured in 14 children. Twenty-two of 34 children with rheumatoid arthritis and none of the 8 controls demonFrom the Children’s Seashore House, Atlantic City, New Jersey; T h e Children’s Hospital of Philadelphia (Connective Tissue Disease Clinic), Philadelphia, Pennsylvania: and the Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania. Supported in part by a grant from the Eastern Pennsylvania Chapter of T h e Arthritis Foundation and the Barsumian Fund. Balu H. Athreya, M.D.: Clinical Director, Children’s Seashore House; Glenda Moser, B.S.: Research Lab Technician, Children’s Seashore House; Henry S. Cecil, M.D.: Associate Professor of Pediatrics, University of Pennsylvania, School of Medicine; Allen R. Myers, M.D.: Chief, Rheumatology Section, University of Pennsylvania, School of Medicine. Address reprints to Dr. Balu H. Athreya, Children’s Seashore House, 4100 Atlantic Avenue, Atlantic City, New Jersey 08401. Submitted for publication August 9, 1974; accepted December 31, 1974. Arthritis and Rheumatism, Vol. 18, No. 4 (July-August 1975)
strated abnormalities of various liver functions at serum salicylate levels between 7.0 and 43 mg%. Three children demonstrated severe abnormalities characterized by marked elevation of SGOT, SGPT, LDH, and AP, prolongation of prothrombin time, and epistaxis. This type of reaction occurred within 5-14 days of initiation of aspirin therapy and occurred at serum salicylate levels between 18 and 43 mg%. Moderate changes in various liver function tests were observed in 19 other children. None of those children who were tested showed prolongation of prothrombin time. T h e serum salicylate level in this group varied between 7.0 and 38.2 mg%. T h e abnormal liver function tests returned to normal in 6 children upon withdrawal of aspirin and in 12 others even when salicylates were continued. Therefore, despite the occurrence of biochemical abnormalities following chronic salicylate therapy, it does not appear to be necessary to discontinue their use except in those children who develop bleeding. Recently there have been reports of salicylateinduced liver injury in juvenile rheumatoid arthritis (JRA) (1,Z). However, these reports do not include adequate control data on children with diseases other than JRA or on liver function abnormalities prior to initiation of salicylate therapy. Therefore, it is difficult to assess whether the abnormal liver function studies reported were due to salicylates or involvement of the liver due to the disease process itself (9-5).
ATHREYA E T AL
348
Table 1. Patients Studied Prospectively for Liver Dysfunction Associated with Salicylate Therapy
of Patients
Number with Abnormal Liver Function Tests on Salicylate Therapy
34
22
Acute cartilaginous necrosis
6
0
Ulcerative colitis with hip disease
2
0
Total Number Juvenile rheumatoid
arthritis
Conversely, there have been many reports of liver disease as part of rheumatoid arthritis without recognition of the possibility that these abnormalities of liver function may be due to drug therapy (3,6,7). I n one report of 7 children with JRA and liver involvement, 3 children appeared to develop abnormalities directly related to salicylate therapy, but this correlation was not recognized in the report (6). Furthermore, the reversibility and/or outcome of chronic salicylate therapy i n JRA is unknown. Therefore a prospective clinical study was initiated of the incidence and extent of liver dysfunction due to salicylates i n children with JRA.
MATERIALS A N D METHODS Thirty-four children attending the Connective Tissue Disease Clinic at the Children’s Hospital of Philadelphia or inpatients of the Children’s Seashore House were followed for periods up to 27 months. Sixteen patients had a monoarticular type of onset, 16 had polyarticular type of onset, and 2 had a febrile type onset. All of them were on aspirin therapy. One was getting weekly injections of gold in addition to aspirin and 1 was on 10 mg of prednisone daily in addition to aspirin. The control group consisted of 6 children with acute cartilaginous necrosis of the hip (ACN) secondary to slipped capital femoral epiphysis and 2 children with ulcerative colitis and hip disease (UC) taking aspirin and followed for 2% months. Unfortunately, there was no indication for more prolonged salicylate therapy in these patients; hence the shorter exposure for the control group. Blood was collected for estimation of liver enzymes, prothrombin time (PT), and salicylate level, at the initial visit and prior to initiation of therapy. If they were already taking salicylates, this treatment was stopped for 3-7 days before blood was drawn for liver function tests. These studies were repeated at least every 3 months and more often in children who were inpatients or in those with marked alterations of liver function tests. Serum glutamic oxaloacetic transaminase (SGOT), serum glutamic pyruvic transaminase (SGPT), lactic dehydrogenase (LDH), and alkaline phosphatase (AP) were measured using standard clinical laboratory procedures.
Table 2. Frequency of Abnormal Liver Function Tests with Chronic Salicylate Administration in Patients with I R A
number with one or more abnormal liver function
TOTAL
SGOT elevated in SGPT elevated in LDH elevated in Alkaline phosphatase elevated in Prothrombin time prolonged in 5’-Nucleotidase elevated in
22 22 14 14 9 3 (of 14 tested) 0 (of 14 tested)
(Normal values for these measurements in our laboratory are given as part of Tables 1-8.) Prothrombin time was measured in 14 of the children using Quick’s method (because these were performed in different laboratories the values are expressed in percentage). 5’-Nudeotidase (5’NT) was measured in 14 children using the method of Dixon and Purdom (8). Total salicylates were measured using Trinder’s method (9). Blood was collected from all the children between 2 and 3 hours after the morning dose of aspirin (except in children taking enteric-coated aspirin in whom blood was collected between 3 and 4 hours after the morning dose). Serum was separated and stored at -70°C until studies were performed.
RESULTS Serum enzyme values that were 2 S.D. above the mean normal and prothrombin time below 70% of normal were considered abnormal. Average value of serum salicylate level in the patient group was 23.3 mgyo (range: 9.0-54 mg%) and i n the control group 15.8 mg% (range: 3.5-29.0 mg%). Twenty-two of the 34 children with JRA showed alterations i n one or more of the liver functions tests (LFT). In contrast, none of the children with ACN and UC showed any alterations of liver functions in serial samples of sera (Table 1). Twenty of these 22 children with abnormalities of liver function were girls. Expressed differently, 2 out of the 8 boys i n the entire group (25%) showed effects of salicylates on liver function tests as against 20 o u t of the 26 girls (77%). Of these 22 patients, 10 had the monoarticular type and 12 had the polyarticular or systemic variety
of JRA. All 22 patients showed elevation of SGOT, 14 showed elevated SGPT, 14 showed elevated LDH, and 9 demonstrated elevation of alkaline phosphatase (Table 2). Three patients had significant prolongation of prothrombin time. 5’NT was not elevated i n any of the sera of 14 patients in which this was assayed. Elevation of LDH was usually noticed first followed
349
ASPIRIN-INDUCED HEPATOTOXICITY I N JRA
Table 3. Patients in Croup 1: Severe Hepatotoxicity
Normal
Serum Salicylate (mg%) 0
SGOT
SGPT
(S-FUnits) 0-28
(S-FUnits) 0-35
Alk P-taw (IU) 30-85
Pro Time
10.2 242 150 230
40 42 64 62
(%) 100
~
Patient DB (1) (2) LF KS
18 31 43
470 700 1580 420
-
180 280 880 190
within 1 week by elevations of SGOT and SGPT. The patients appeared to fall into two distinct groups. The first group consisted of 3 patients (all of them girls) who showed marked elevations of all the enzymes tested, prolongation of prothrombin time, and epistaxis severe enough to require hospitalization (Table 3). The serum salicylate level when these alterations occurred varied between 18 and 43 mga/, in 2 patients. In the third patient, serum salicylate level was not obtained at the time of the admission. All of them recovered completely both clinically and biochemically upon withdrawal of aspirin. Australiaantigen was looked for in 2 patients using a radioimmunoassay and none was found. Liver biopsy performed during peak biochemical changes in 1 of these patients showed minimal periportal infiltration with lymphocytes. An example of such a patient was a 10-year-old black girl with JRA (Table 4). Before aspirin therapy, prothrombin time and levels of SGOT and SGPT were normal. Alkaline phosphatase was 103 U. Two weeks after aspirin therapy was initiated serum salicylate level was 43 mg%. At this time SGOT and SGPT were markedly elevated and prothrombin time was 64%. She had severe epistaxis and aspirin was stopped. Ten days later her SGOT and SGPT values were still high. Alkaline phosphatase level was normal. The prothrombin time was 70y0. Salicylate therapy was reinstituted at a lesser dose on more than one occasion but was not maintained because of persistent abnormalities of liver function. Table 4. Profile
Normal Date 2-20-73 3-5-73 3-14-73 4-26-73
Serum Salicylate (mg%) 0 0 43 22 6.5
SGOT (S-F Units)
0-28
25 1580 174 43
of
The second group (Table 5 ) consisted of 19 patients who showed moderate elevations of one or more enzymes. Prothrombin time was measured at least once in 11 of these patients and was found to be normal. None of these patients had epistaxis. The serum salicylate level at which these abnormalities occurred varied between 7.0 and 38.2 me%. LDH elevation, as noted above, preceded elevations of other enzymes by 7-1 0 days. These elevations were documented on more than one occasion with alterations of different enzymes at different times. Australia-antigen was not demonstrated in 5 of the patients from this group, in whom it was looked for. An example of this group is patient DL. As can be seen in Table 6, the results of liver function tests showed return of values toward normal even when aspirin was continued. Nine of the children developed liver function abnormalities within the first month after starting aspirin therapy. Seven patients developed abnormalities of liver function between 1 and 3 months; 2 patients between 3 and 6 months; and 4 patients between 7 and 18 months after initiation of aspirin therapy (Table 7). Liver function returned to normal on withdrawal of aspirin in 6 of the 22 children. Twelve others showed fluctuation in enzyme levels without any deterioration even on continuing therapy at the same level. It is too early to evaluate the status on 4 others (Table 8). In addition, efforts were made to define the Patient LF (Group 1) SGPT
(S-FUnits) &35
5 880 250 17.5
Alk P-tase (IU) 17-83
ProTime
103
100 64 70 80
150
67.5 83
(%)
100
LDH (Iv) 2478
720 210 150
350
ATHREYA E T AL
Table 5. Patients in Group 2 : Moderate Hepatotoxicity
Normal Patient DL LM
cv
DB JD JB CB SL
NP MB RD RD SAM RM KP PS
ss DW cc
Serum Salicylate (mgY0) 0
SGOT
SGPT (S-F Units) 0-35
(S-F Units)
0-28
145 186 58 165 100 74 55 97 58 40 70 97 52 318 40 89 46 46 116
30.2
-
17.0 27.4 20.0 17.0 19.7 38.2 11.2 20.0 11.0 32.0 7.0 30.0 24.0 23.3 26.8 23 28
25 200 53 135 124 39 43 43 21 14.5 14 90 34.5 600 17.5 57 17.5 21 76.5
'4Ik P-tase (IU) 17-83
51 67.5 79.5 38 92 55 45 101 45 89 57 75 114 64 45 80 71.5 92
Pro Time
(70)
LD H
(W
100
24-78
100 -
-
100 100 100 90 100 97
-
100
-
100 90 100
22
101 160 76 80
-
86 47 76 75 95 121 134 107 85 58 82 110
Note: Representative sample of each patient.
in the 8 patients in whom it was looked for.
tissue origin of LDH and AP in a group of patients with elevated levels of these enzymes. Because the liver fraction of LDH is heat labile, heating the serum for 30 minutes at 65OC will result in reduction of LDH levels by 55% or more of unheated serum, if the elevation is due primarily to liver fraction (10). On the other hand, heating serum at 57°C for 15 minutes will not affect the AP levels if i t is from the liver but will affect the bone fraction (11). These studies performed on the sera of 9 patients with elevated LDH levels and 4 patients with elevated AP levels demonstrated that both LDH and AP elevations were due to liver fractions. Addition of aspirin in vitro to serum samples (at concentrations of 25 mg%) did not alter or interfere with the measurement of liver enzymes. Finally, because the adverse effects of salicylates on the liver might be due to a hypersensitivity reaction eosinophilia might be expected to occur. None was found
DISCUSSION When this clinical study was begun in 1970, there were very few publications on the effects of salicylates on the liver i n rheumatic diseases (12-14). T h e earliest one described the effect of salicylates on the liver in patients treated for acute rheumatic fever (12). Since then there have been several reports, including one of our own, on salicylate-induced hepatoxicity in rheumatoid arthritis (1,2,15). Salicylates have been associated with hepatic injury in systemic lupus erythematosus also (16,17). T h e present study confirms the occurrence of hepatic dysfunction associated with salicylate therapy in JRA and provides information relative to the incidence of the problem and consequences of continued therapy. T h e incidence of liver function abnormalities in 22 out of 34 children (65%) with J R A confirms and
Table 6. Profile of Patient DL (Group 2)
Normal ~~
Serum Salicylate
SGOT
(mg70)
(S-F Units)
0
0-28
SGPT (S-F Units) 0-35
Alk P-tase (1")
LDH (IU) 2478
51 86 83 59
117 97 80
17-83
~~
Date 5-25-72 4-6-73 6-1-73 9-7-73
30.2 33.8 38.8 22
145 130 66 34
25 47 39 37
ASPIRIN-INDUCED HEPATOTOXICITY IN JRA
Table 7. T i m e Znterual from Initiation of Aspirin Therapy to Development of Hepatic Function Abnormalities in I R A Number of Patients With 1 month Between 1 and 3 months Between 3 and 6 months Six months or more
55 1
Table 8. Outcome of Salicylate-Znduced Liver Dysfunction Return of values to normal on stopping aspirin Return of values toward or to normal even when high dose of aspirin was maintained Persistence of abnormalities
6
12 4 Total 22
Total 22
extends the earlier findings of Russell et a1 (1) and Rich and Johnson (2). Russell et a1 found abnormalities of serum transaminase in 8 of 32 patients. Rich and Johnson reported 6 patients with abnormal liver functions. All of them had elevated SGOT, SGPT, 4 had elevated LDH, 4 had elevated AP, 1 had increased retention of BSP, and 1 had elevated serum bilirubin. In the study by Russell et a1 (1) serum salicylate level was over 35 mg% in all but 1 patient when transaminase levels were elevated. In the study by Rich and Johnson (2) all patients had serum salicylate levels over 25 mg% when liver functions abnormalities were noted. However in this study liver function abnormalities were noted at serum salicylate levels as low as 7.0 mg% (Table 5). One patient, reported earlier in detail, developed marked prolongation of prothrombin time and elevated values of SGOT, SGPT, and AP at a serum salicylate level of 18 mg%. Therefore it might be concluded that the effect of aspirin on the liver is not entirely dose-related. None of the control group of children with ulcerative colitis or acute cartilaginous necrosis showed such abnormalities. Though 2 of them had low serum salicylate levels, the others had serum levels above 15 mg% (maximum: 29 mg%). This finding combined with the observation that liver function abnormalities associated with aspirin therapy have been noted in acute rheumatic fever (12), JRA (1,2,15), and SLE (16,17) suggests that some feature of these conditions, the prolonged high doses of aspirin used in these conditions, or a combination of these is responsible for the demonstrated liver function abnormalities. This reaction does not appear to be one of hypersensitivity type for the following reasons: a) There was no fixed latent period. One girl developed changes in liver function 4 days after therapy was started and 1 developed these changes 18 months after salicylate therapy was started. b) None of the patients had rash. c) Eosinophilia, when examined for before
or during the peak biochemical changes, was not seen (though Rich and Johnson found eosinophilia to be a good indicator of occurrence of liver injury). T h e very high levels of serum transaminases demonstrated in this and other studies suggests that the injury is hepatocellular. Both in this study and in earlier studies, elevation of alkalkine phosphatase was noted only in a few children and was mild and hard to interpret because of normal variations in growing children and the influence of joint disease and osteoporosis on alkaline phosphatase values (18, 19). In growing children the liver component of serum AP is masked by osseous factors. Therefore, alanine aminotransferase and 5’-nucleotidase, which are not age dependent, have been considered more sensitive indicators of liver damage (18,ZO). In this study, none of the patients tested showed elevated 5’NT levels. Even the 1 patient with very high level of SGOT, SGPT, and LDH had normal 5’NT value. Therefore, it appears that the injury is of the hepatocellular type. However, this is only biochemical injury. Liver biopsy in 1 patient in the group, at the time of peak biochemical abnormalities, showed minimal abnormalities (periportal lymphocytic infiltration). Rich and Johnson also demonstrated similar findings in 2 patients in whom liver biopsies were performed (2). At present, it appears that the abnormal liver functions are totally reversible upon withdrawal of aspirin. In 6 of the 22 children all liver function abnormalities were reversed upon withdrawal of aspirin. In 12 others the values returned to normal even when aspirin therapy was continued without reduction of dosage (Table 8). One important relevant observation in this study is the separation of these 22 children with abnormal liver function into two distinct groups: Group 1 included 3 children with marked elevations of serum enzymes, prolongation of prothrombin time, and epistaxis. Obviously, these children cannot tolerate aspirin. Group 2 included 19 children with mild to moderate elevations of serum enzymes, normal prothrombin time, and no bleeding episodes. In these children, salicylate therapy was continued even in the presence
ATHREYA E T AL
of abnormal liver function tests, as long as the prothrombin time remained normal. As indicated earlier, 12 showed return of liver function to normal even when salicylate therapy was continued. In the first group, therapy was initiated with lower doses of aspirin to see i f antirheumatic activity could be achieved without affecting the liver. This result was not possible. I n the second group, reduction of dose appears unnecessary. Another interesting observation in this study is the association between the sex of the patient and the occurrence of liver function abnormalities. Only 2 of the 22 with liver function abnormalities were boys. This number is an incidence of 25% in boys (2 of 8 boys included in the study) and 77% in girls (20 of 26 girls). One possible explanation is a difference in the metabolism of aspirin between boys and girls. However, the number of boys in our study is too small for firm conclusions. This point is currently being further investigated. Therefore it appears that liver function abnormalities are very common in children with JRA on chronic salicylate therapy. Because in some patients these can be of serious consequence, liver function studies including prothrombin time should be performed before salicylate therapy is initiated and at regular intervals thereafter. Except for those patients with bleeding disorders and severe liver dysfunction (Group 1) cessation of salicylate therapy is not necessary. In fact, the return of liver functions to normal in most patients while on chronic salicylate therapy provides reassurance for the continued use of this valuable therapeutic agent in the treatment of juvenile rheumatoid arthritis.
REFERENCES Russell AS, Sturge RA, and Smith MA: Serum transaminases during salicylate therapy. Br Med J 2428429,1971 Rich RR, Johnson JA: Salicylate hepatotoxicity in patients with juvenile rheumatoid arthritis. Arthritis Rheum 16:1-9, 1973 Darby PW: Liver function tests in rheumatoid arthritis. J Clin Pathol9: 153-156, 1956
4. Subclinical hepatitis in connective tissue diseases. Lancet 1: 1095-1096, 1970 (editorial) 5. Schaller J: T h e liver and arthritis. J Pediatr 79:13% 141, 1971 6. Kornreich H, Malouf NN, Hanson V: Acute hepatic dysfunction in juvenile rheumatoid arthritis. J Pediatr 79: 27-35, 1971 7. Cockel R, Kendall MJ, Becker JF, et al: Serum biochemical values in rheumatoid disease. Ann Rheum Dis 30: 166-170, 1971 8. Dixon TF, Purdom MJ: Serum 5’nucleotidase. J Clin Pathol 7:341, 1954 9. Trinder P: Rapid determination of salicylate in biological fluids. Biochem J 57:301, 1954 10. Bauer JD, Ackerman PG, Tor0 G: Bray’s Clinical Laboratory Methods. Seventh edition, St. Louis, CV Mosby CO, 1968, pp 397-398 11. Posen S . Neale FC, Clubb JS: Heat inactivation in the study of human alkaline phosphatase. Ann Inter! Med 62:1233-1243.1965 12. Manso C, Tranata A, Nydick I: Effect of aspirin administration on serum glutamic oxaloacetic and glutamic pyruvic transaminases in children. Proc SOCExp Biol Med 93:84-87, 1956 13, Chalmers TM, Kellgren JH, Platt DS: Evaluation in man of Fenclozic Acid (I.C.I. 54, 450:Myalex), a new anti-inflammatory agent. 11. Clinical trials in patients with rheumatoid arthritis. Ann Rheum Dis 28:595-601, 1969 14. Iancu T: Serum transaminase and salicylate therapy. Br Med J 2167, 1972 15. Athreya BH, Gorske AL, Myers AR: Aspirin-induced abnormalities of liver dysfunction. Am J Dis Child 126:638-641, 1973 16. Seaman WE, Ishak KG, Plotz PH: Aspirin-induced hepatotoxicity in patients with systemic lupus erythematosus. Ann Intern Med 8O:l-8, 1974 17. Wolfe JD, Metzger AL, Goldstein RC: Aspirin hepatitis. Ann Intern Med 80:74-76, 1974 18. Alkaline phosphatase values in children. J Pediatr 82: 361-362, 1973 (editorial) 19. Posen S: Alkaline phosphatase. Ann Intern Med 67: 185-2021, 1967 20. Kattwinkel J, Taussig LM, Statland BE, et al: The effects of age on alkaline phosphatase and other serologic liver function tests in normal subjects and patients with cystic fibrosis. J Pediatr 82:234-242, 1973
ASPIRIN-INDUCED HEPATOTOXICITY IN JUVENILE RHEUMATOID ARTHRITIS A PROSPECTIVE STUDY BALU H. ATHREYA, GLENDA MOSER, HENRY S. CECIL, and ALLEN R. MYERS
Recent reports of and our own experience with biochemical alterations of liver function secondary to salicylate therapy stimulated this prospective study. Thirty-four children with juvenile rheumatoid arthritis, 6 children with acute cartilagenous necrosis of the hip following slipped capital femoral epiphysis, and 2 children with ulcerative colitis and hip disease who were on salicylates were followed over a period of 1-27 months with serial determinations of serum glutamic oxaloacetic transaminase (SGOT), serum glutamic pyruvic transaminase (SGPT), lactic dehydrogenase (LDH), alkaline phosphatase (AP), bilirubin, and serum salicylate. Prothrombin time was measured in 14 children. Twenty-two of 34 children with rheumatoid arthritis and none of the 8 controls demonFrom the Children’s Seashore House, Atlantic City, New Jersey; T h e Children’s Hospital of Philadelphia (Connective Tissue Disease Clinic), Philadelphia, Pennsylvania: and the Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania. Supported in part by a grant from the Eastern Pennsylvania Chapter of T h e Arthritis Foundation and the Barsumian Fund. Balu H. Athreya, M.D.: Clinical Director, Children’s Seashore House; Glenda Moser, B.S.: Research Lab Technician, Children’s Seashore House; Henry S. Cecil, M.D.: Associate Professor of Pediatrics, University of Pennsylvania, School of Medicine; Allen R. Myers, M.D.: Chief, Rheumatology Section, University of Pennsylvania, School of Medicine. Address reprints to Dr. Balu H. Athreya, Children’s Seashore House, 4100 Atlantic Avenue, Atlantic City, New Jersey 08401. Submitted for publication August 9, 1974; accepted December 31, 1974. Arthritis and Rheumatism, Vol. 18, No. 4 (July-August 1975)
strated abnormalities of various liver functions at serum salicylate levels between 7.0 and 43 mg%. Three children demonstrated severe abnormalities characterized by marked elevation of SGOT, SGPT, LDH, and AP, prolongation of prothrombin time, and epistaxis. This type of reaction occurred within 5-14 days of initiation of aspirin therapy and occurred at serum salicylate levels between 18 and 43 mg%. Moderate changes in various liver function tests were observed in 19 other children. None of those children who were tested showed prolongation of prothrombin time. T h e serum salicylate level in this group varied between 7.0 and 38.2 mg%. T h e abnormal liver function tests returned to normal in 6 children upon withdrawal of aspirin and in 12 others even when salicylates were continued. Therefore, despite the occurrence of biochemical abnormalities following chronic salicylate therapy, it does not appear to be necessary to discontinue their use except in those children who develop bleeding. Recently there have been reports of salicylateinduced liver injury in juvenile rheumatoid arthritis (JRA) (1,Z). However, these reports do not include adequate control data on children with diseases other than JRA or on liver function abnormalities prior to initiation of salicylate therapy. Therefore, it is difficult to assess whether the abnormal liver function studies reported were due to salicylates or involvement of the liver due to the disease process itself (9-5).
ATHREYA E T AL
348
Table 1. Patients Studied Prospectively for Liver Dysfunction Associated with Salicylate Therapy
of Patients
Number with Abnormal Liver Function Tests on Salicylate Therapy
34
22
Acute cartilaginous necrosis
6
0
Ulcerative colitis with hip disease
2
0
Total Number Juvenile rheumatoid
arthritis
Conversely, there have been many reports of liver disease as part of rheumatoid arthritis without recognition of the possibility that these abnormalities of liver function may be due to drug therapy (3,6,7). I n one report of 7 children with JRA and liver involvement, 3 children appeared to develop abnormalities directly related to salicylate therapy, but this correlation was not recognized in the report (6). Furthermore, the reversibility and/or outcome of chronic salicylate therapy i n JRA is unknown. Therefore a prospective clinical study was initiated of the incidence and extent of liver dysfunction due to salicylates i n children with JRA.
MATERIALS A N D METHODS Thirty-four children attending the Connective Tissue Disease Clinic at the Children’s Hospital of Philadelphia or inpatients of the Children’s Seashore House were followed for periods up to 27 months. Sixteen patients had a monoarticular type of onset, 16 had polyarticular type of onset, and 2 had a febrile type onset. All of them were on aspirin therapy. One was getting weekly injections of gold in addition to aspirin and 1 was on 10 mg of prednisone daily in addition to aspirin. The control group consisted of 6 children with acute cartilaginous necrosis of the hip (ACN) secondary to slipped capital femoral epiphysis and 2 children with ulcerative colitis and hip disease (UC) taking aspirin and followed for 2% months. Unfortunately, there was no indication for more prolonged salicylate therapy in these patients; hence the shorter exposure for the control group. Blood was collected for estimation of liver enzymes, prothrombin time (PT), and salicylate level, at the initial visit and prior to initiation of therapy. If they were already taking salicylates, this treatment was stopped for 3-7 days before blood was drawn for liver function tests. These studies were repeated at least every 3 months and more often in children who were inpatients or in those with marked alterations of liver function tests. Serum glutamic oxaloacetic transaminase (SGOT), serum glutamic pyruvic transaminase (SGPT), lactic dehydrogenase (LDH), and alkaline phosphatase (AP) were measured using standard clinical laboratory procedures.
Table 2. Frequency of Abnormal Liver Function Tests with Chronic Salicylate Administration in Patients with I R A
number with one or more abnormal liver function
TOTAL
SGOT elevated in SGPT elevated in LDH elevated in Alkaline phosphatase elevated in Prothrombin time prolonged in 5’-Nucleotidase elevated in
22 22 14 14 9 3 (of 14 tested) 0 (of 14 tested)
(Normal values for these measurements in our laboratory are given as part of Tables 1-8.) Prothrombin time was measured in 14 of the children using Quick’s method (because these were performed in different laboratories the values are expressed in percentage). 5’-Nudeotidase (5’NT) was measured in 14 children using the method of Dixon and Purdom (8). Total salicylates were measured using Trinder’s method (9). Blood was collected from all the children between 2 and 3 hours after the morning dose of aspirin (except in children taking enteric-coated aspirin in whom blood was collected between 3 and 4 hours after the morning dose). Serum was separated and stored at -70°C until studies were performed.
RESULTS Serum enzyme values that were 2 S.D. above the mean normal and prothrombin time below 70% of normal were considered abnormal. Average value of serum salicylate level in the patient group was 23.3 mgyo (range: 9.0-54 mg%) and i n the control group 15.8 mg% (range: 3.5-29.0 mg%). Twenty-two of the 34 children with JRA showed alterations i n one or more of the liver functions tests (LFT). In contrast, none of the children with ACN and UC showed any alterations of liver functions in serial samples of sera (Table 1). Twenty of these 22 children with abnormalities of liver function were girls. Expressed differently, 2 out of the 8 boys i n the entire group (25%) showed effects of salicylates on liver function tests as against 20 o u t of the 26 girls (77%). Of these 22 patients, 10 had the monoarticular type and 12 had the polyarticular or systemic variety
of JRA. All 22 patients showed elevation of SGOT, 14 showed elevated SGPT, 14 showed elevated LDH, and 9 demonstrated elevation of alkaline phosphatase (Table 2). Three patients had significant prolongation of prothrombin time. 5’NT was not elevated i n any of the sera of 14 patients in which this was assayed. Elevation of LDH was usually noticed first followed
349
ASPIRIN-INDUCED HEPATOTOXICITY I N JRA
Table 3. Patients in Croup 1: Severe Hepatotoxicity
Normal
Serum Salicylate (mg%) 0
SGOT
SGPT
(S-FUnits) 0-28
(S-FUnits) 0-35
Alk P-taw (IU) 30-85
Pro Time
10.2 242 150 230
40 42 64 62
(%) 100
~
Patient DB (1) (2) LF KS
18 31 43
470 700 1580 420
-
180 280 880 190
within 1 week by elevations of SGOT and SGPT. The patients appeared to fall into two distinct groups. The first group consisted of 3 patients (all of them girls) who showed marked elevations of all the enzymes tested, prolongation of prothrombin time, and epistaxis severe enough to require hospitalization (Table 3). The serum salicylate level when these alterations occurred varied between 18 and 43 mga/, in 2 patients. In the third patient, serum salicylate level was not obtained at the time of the admission. All of them recovered completely both clinically and biochemically upon withdrawal of aspirin. Australiaantigen was looked for in 2 patients using a radioimmunoassay and none was found. Liver biopsy performed during peak biochemical changes in 1 of these patients showed minimal periportal infiltration with lymphocytes. An example of such a patient was a 10-year-old black girl with JRA (Table 4). Before aspirin therapy, prothrombin time and levels of SGOT and SGPT were normal. Alkaline phosphatase was 103 U. Two weeks after aspirin therapy was initiated serum salicylate level was 43 mg%. At this time SGOT and SGPT were markedly elevated and prothrombin time was 64%. She had severe epistaxis and aspirin was stopped. Ten days later her SGOT and SGPT values were still high. Alkaline phosphatase level was normal. The prothrombin time was 70y0. Salicylate therapy was reinstituted at a lesser dose on more than one occasion but was not maintained because of persistent abnormalities of liver function. Table 4. Profile
Normal Date 2-20-73 3-5-73 3-14-73 4-26-73
Serum Salicylate (mg%) 0 0 43 22 6.5
SGOT (S-F Units)
0-28
25 1580 174 43
of
The second group (Table 5 ) consisted of 19 patients who showed moderate elevations of one or more enzymes. Prothrombin time was measured at least once in 11 of these patients and was found to be normal. None of these patients had epistaxis. The serum salicylate level at which these abnormalities occurred varied between 7.0 and 38.2 me%. LDH elevation, as noted above, preceded elevations of other enzymes by 7-1 0 days. These elevations were documented on more than one occasion with alterations of different enzymes at different times. Australia-antigen was not demonstrated in 5 of the patients from this group, in whom it was looked for. An example of this group is patient DL. As can be seen in Table 6, the results of liver function tests showed return of values toward normal even when aspirin was continued. Nine of the children developed liver function abnormalities within the first month after starting aspirin therapy. Seven patients developed abnormalities of liver function between 1 and 3 months; 2 patients between 3 and 6 months; and 4 patients between 7 and 18 months after initiation of aspirin therapy (Table 7). Liver function returned to normal on withdrawal of aspirin in 6 of the 22 children. Twelve others showed fluctuation in enzyme levels without any deterioration even on continuing therapy at the same level. It is too early to evaluate the status on 4 others (Table 8). In addition, efforts were made to define the Patient LF (Group 1) SGPT
(S-FUnits) &35
5 880 250 17.5
Alk P-tase (IU) 17-83
ProTime
103
100 64 70 80
150
67.5 83
(%)
100
LDH (Iv) 2478
720 210 150
350
ATHREYA E T AL
Table 5. Patients in Group 2 : Moderate Hepatotoxicity
Normal Patient DL LM
cv
DB JD JB CB SL
NP MB RD RD SAM RM KP PS
ss DW cc
Serum Salicylate (mgY0) 0
SGOT
SGPT (S-F Units) 0-35
(S-F Units)
0-28
145 186 58 165 100 74 55 97 58 40 70 97 52 318 40 89 46 46 116
30.2
-
17.0 27.4 20.0 17.0 19.7 38.2 11.2 20.0 11.0 32.0 7.0 30.0 24.0 23.3 26.8 23 28
25 200 53 135 124 39 43 43 21 14.5 14 90 34.5 600 17.5 57 17.5 21 76.5
'4Ik P-tase (IU) 17-83
51 67.5 79.5 38 92 55 45 101 45 89 57 75 114 64 45 80 71.5 92
Pro Time
(70)
LD H
(W
100
24-78
100 -
-
100 100 100 90 100 97
-
100
-
100 90 100
22
101 160 76 80
-
86 47 76 75 95 121 134 107 85 58 82 110
Note: Representative sample of each patient.
in the 8 patients in whom it was looked for.
tissue origin of LDH and AP in a group of patients with elevated levels of these enzymes. Because the liver fraction of LDH is heat labile, heating the serum for 30 minutes at 65OC will result in reduction of LDH levels by 55% or more of unheated serum, if the elevation is due primarily to liver fraction (10). On the other hand, heating serum at 57°C for 15 minutes will not affect the AP levels if i t is from the liver but will affect the bone fraction (11). These studies performed on the sera of 9 patients with elevated LDH levels and 4 patients with elevated AP levels demonstrated that both LDH and AP elevations were due to liver fractions. Addition of aspirin in vitro to serum samples (at concentrations of 25 mg%) did not alter or interfere with the measurement of liver enzymes. Finally, because the adverse effects of salicylates on the liver might be due to a hypersensitivity reaction eosinophilia might be expected to occur. None was found
DISCUSSION When this clinical study was begun in 1970, there were very few publications on the effects of salicylates on the liver i n rheumatic diseases (12-14). T h e earliest one described the effect of salicylates on the liver in patients treated for acute rheumatic fever (12). Since then there have been several reports, including one of our own, on salicylate-induced hepatoxicity in rheumatoid arthritis (1,2,15). Salicylates have been associated with hepatic injury in systemic lupus erythematosus also (16,17). T h e present study confirms the occurrence of hepatic dysfunction associated with salicylate therapy in JRA and provides information relative to the incidence of the problem and consequences of continued therapy. T h e incidence of liver function abnormalities in 22 out of 34 children (65%) with J R A confirms and
Table 6. Profile of Patient DL (Group 2)
Normal ~~
Serum Salicylate
SGOT
(mg70)
(S-F Units)
0
0-28
SGPT (S-F Units) 0-35
Alk P-tase (1")
LDH (IU) 2478
51 86 83 59
117 97 80
17-83
~~
Date 5-25-72 4-6-73 6-1-73 9-7-73
30.2 33.8 38.8 22
145 130 66 34
25 47 39 37
ASPIRIN-INDUCED HEPATOTOXICITY IN JRA
Table 7. T i m e Znterual from Initiation of Aspirin Therapy to Development of Hepatic Function Abnormalities in I R A Number of Patients With 1 month Between 1 and 3 months Between 3 and 6 months Six months or more
55 1
Table 8. Outcome of Salicylate-Znduced Liver Dysfunction Return of values to normal on stopping aspirin Return of values toward or to normal even when high dose of aspirin was maintained Persistence of abnormalities
6
12 4 Total 22
Total 22
extends the earlier findings of Russell et a1 (1) and Rich and Johnson (2). Russell et a1 found abnormalities of serum transaminase in 8 of 32 patients. Rich and Johnson reported 6 patients with abnormal liver functions. All of them had elevated SGOT, SGPT, 4 had elevated LDH, 4 had elevated AP, 1 had increased retention of BSP, and 1 had elevated serum bilirubin. In the study by Russell et a1 (1) serum salicylate level was over 35 mg% in all but 1 patient when transaminase levels were elevated. In the study by Rich and Johnson (2) all patients had serum salicylate levels over 25 mg% when liver functions abnormalities were noted. However in this study liver function abnormalities were noted at serum salicylate levels as low as 7.0 mg% (Table 5). One patient, reported earlier in detail, developed marked prolongation of prothrombin time and elevated values of SGOT, SGPT, and AP at a serum salicylate level of 18 mg%. Therefore it might be concluded that the effect of aspirin on the liver is not entirely dose-related. None of the control group of children with ulcerative colitis or acute cartilaginous necrosis showed such abnormalities. Though 2 of them had low serum salicylate levels, the others had serum levels above 15 mg% (maximum: 29 mg%). This finding combined with the observation that liver function abnormalities associated with aspirin therapy have been noted in acute rheumatic fever (12), JRA (1,2,15), and SLE (16,17) suggests that some feature of these conditions, the prolonged high doses of aspirin used in these conditions, or a combination of these is responsible for the demonstrated liver function abnormalities. This reaction does not appear to be one of hypersensitivity type for the following reasons: a) There was no fixed latent period. One girl developed changes in liver function 4 days after therapy was started and 1 developed these changes 18 months after salicylate therapy was started. b) None of the patients had rash. c) Eosinophilia, when examined for before
or during the peak biochemical changes, was not seen (though Rich and Johnson found eosinophilia to be a good indicator of occurrence of liver injury). T h e very high levels of serum transaminases demonstrated in this and other studies suggests that the injury is hepatocellular. Both in this study and in earlier studies, elevation of alkalkine phosphatase was noted only in a few children and was mild and hard to interpret because of normal variations in growing children and the influence of joint disease and osteoporosis on alkaline phosphatase values (18, 19). In growing children the liver component of serum AP is masked by osseous factors. Therefore, alanine aminotransferase and 5’-nucleotidase, which are not age dependent, have been considered more sensitive indicators of liver damage (18,ZO). In this study, none of the patients tested showed elevated 5’NT levels. Even the 1 patient with very high level of SGOT, SGPT, and LDH had normal 5’NT value. Therefore, it appears that the injury is of the hepatocellular type. However, this is only biochemical injury. Liver biopsy in 1 patient in the group, at the time of peak biochemical abnormalities, showed minimal abnormalities (periportal lymphocytic infiltration). Rich and Johnson also demonstrated similar findings in 2 patients in whom liver biopsies were performed (2). At present, it appears that the abnormal liver functions are totally reversible upon withdrawal of aspirin. In 6 of the 22 children all liver function abnormalities were reversed upon withdrawal of aspirin. In 12 others the values returned to normal even when aspirin therapy was continued without reduction of dosage (Table 8). One important relevant observation in this study is the separation of these 22 children with abnormal liver function into two distinct groups: Group 1 included 3 children with marked elevations of serum enzymes, prolongation of prothrombin time, and epistaxis. Obviously, these children cannot tolerate aspirin. Group 2 included 19 children with mild to moderate elevations of serum enzymes, normal prothrombin time, and no bleeding episodes. In these children, salicylate therapy was continued even in the presence
ATHREYA E T AL
of abnormal liver function tests, as long as the prothrombin time remained normal. As indicated earlier, 12 showed return of liver function to normal even when salicylate therapy was continued. In the first group, therapy was initiated with lower doses of aspirin to see i f antirheumatic activity could be achieved without affecting the liver. This result was not possible. I n the second group, reduction of dose appears unnecessary. Another interesting observation in this study is the association between the sex of the patient and the occurrence of liver function abnormalities. Only 2 of the 22 with liver function abnormalities were boys. This number is an incidence of 25% in boys (2 of 8 boys included in the study) and 77% in girls (20 of 26 girls). One possible explanation is a difference in the metabolism of aspirin between boys and girls. However, the number of boys in our study is too small for firm conclusions. This point is currently being further investigated. Therefore it appears that liver function abnormalities are very common in children with JRA on chronic salicylate therapy. Because in some patients these can be of serious consequence, liver function studies including prothrombin time should be performed before salicylate therapy is initiated and at regular intervals thereafter. Except for those patients with bleeding disorders and severe liver dysfunction (Group 1) cessation of salicylate therapy is not necessary. In fact, the return of liver functions to normal in most patients while on chronic salicylate therapy provides reassurance for the continued use of this valuable therapeutic agent in the treatment of juvenile rheumatoid arthritis.
REFERENCES Russell AS, Sturge RA, and Smith MA: Serum transaminases during salicylate therapy. Br Med J 2428429,1971 Rich RR, Johnson JA: Salicylate hepatotoxicity in patients with juvenile rheumatoid arthritis. Arthritis Rheum 16:1-9, 1973 Darby PW: Liver function tests in rheumatoid arthritis. J Clin Pathol9: 153-156, 1956
4. Subclinical hepatitis in connective tissue diseases. Lancet 1: 1095-1096, 1970 (editorial) 5. Schaller J: T h e liver and arthritis. J Pediatr 79:13% 141, 1971 6. Kornreich H, Malouf NN, Hanson V: Acute hepatic dysfunction in juvenile rheumatoid arthritis. J Pediatr 79: 27-35, 1971 7. Cockel R, Kendall MJ, Becker JF, et al: Serum biochemical values in rheumatoid disease. Ann Rheum Dis 30: 166-170, 1971 8. Dixon TF, Purdom MJ: Serum 5’nucleotidase. J Clin Pathol 7:341, 1954 9. Trinder P: Rapid determination of salicylate in biological fluids. Biochem J 57:301, 1954 10. Bauer JD, Ackerman PG, Tor0 G: Bray’s Clinical Laboratory Methods. Seventh edition, St. Louis, CV Mosby CO, 1968, pp 397-398 11. Posen S . Neale FC, Clubb JS: Heat inactivation in the study of human alkaline phosphatase. Ann Inter! Med 62:1233-1243.1965 12. Manso C, Tranata A, Nydick I: Effect of aspirin administration on serum glutamic oxaloacetic and glutamic pyruvic transaminases in children. Proc SOCExp Biol Med 93:84-87, 1956 13, Chalmers TM, Kellgren JH, Platt DS: Evaluation in man of Fenclozic Acid (I.C.I. 54, 450:Myalex), a new anti-inflammatory agent. 11. Clinical trials in patients with rheumatoid arthritis. Ann Rheum Dis 28:595-601, 1969 14. Iancu T: Serum transaminase and salicylate therapy. Br Med J 2167, 1972 15. Athreya BH, Gorske AL, Myers AR: Aspirin-induced abnormalities of liver dysfunction. Am J Dis Child 126:638-641, 1973 16. Seaman WE, Ishak KG, Plotz PH: Aspirin-induced hepatotoxicity in patients with systemic lupus erythematosus. Ann Intern Med 8O:l-8, 1974 17. Wolfe JD, Metzger AL, Goldstein RC: Aspirin hepatitis. Ann Intern Med 80:74-76, 1974 18. Alkaline phosphatase values in children. J Pediatr 82: 361-362, 1973 (editorial) 19. Posen S: Alkaline phosphatase. Ann Intern Med 67: 185-2021, 1967 20. Kattwinkel J, Taussig LM, Statland BE, et al: The effects of age on alkaline phosphatase and other serologic liver function tests in normal subjects and patients with cystic fibrosis. J Pediatr 82:234-242, 1973
