Methods

Top Methods Results Discussion Author & Article Info References

Patients

Between July 1990 and August 1991, 95 patients with hyperthyroidism were screened for serum aspartate aminotransferase (AST) (normal, <0.57 µkat/L), alanine aminotransferase (ALT) (normal, <0.60 µkat/L), bilirubin (normal, <22.2 µmol/L), and alkaline phosphatase (ALP) (normal, <1.57 µkat/L) levels as well as for hepatitis B surface antigen (HBsAg) and antibodies to HCV (anti-HCV) before antithyroid therapy. Of these 95 patients, 60 [63%] showed normal AST and ALT levels; these patients formed the basis of our study. The diagnosis of hyperthyroidism was made based on 1) typical symptoms (asthenia, nervousness, heat intolerance, palpitation, and weight loss), and signs [thyroid enlargement, hand tremor, and eye signs]; 2) high serum thyroxine (T₄), triiodothyronine (T₃), and free T₄ levels and low thyroid-stimulating hormone [TSH] levels; and 3) increased Technetium-99m (Technetium-99m) thyroid uptake with diffuse uptake noted on a thyroid scan. No patient had evidence of cardiovascular complications. Patients who had previously been treated for hyperthyroidism were excluded from this study.

Treatment

All patients with hyperthyroidism received 300 mg/d oral propylthiouracil (Procil, Nysco Co. LTD., Taipei, Taiwan) for 2 months. The dose of propylthiouracil was subsequently reduced to 100 to 150 mg/d according to the results of follow-up thyroid function tests and was maintained at 100 mg/d when a euthyroid state was achieved.

Follow-up

Serum AST, ALT, ALP, and bilirubin studies were scheduled in all patients 2 and 5 months after the start of propylthiouracil therapy. After a serum AST or ALT elevation was detected, patients were monitored closely with clinical evaluation and weekly serum AST, ALT, ALP, and bilirubin measurements until improvements were seen. It was determined before the study that PTU was to be discontinued promptly if clinical symptoms of hepatitis or hyperbilirubinemia developed. Serologic markers of hepatitis A virus (HAV), B virus (HBV), delta virus (HDV), and HCV as well as anti-nuclear antibody and anti-smooth muscle antibody were also studied when appropriate. Liver biopsy was routinely advised when abnormal AST or ALT levels were detected but was done in only three patients who gave written, informed consent.

Laboratory Methods

Tests of serum AST, ALT, ALP, and bilirubin were done in the clinical pathology laboratories of Chang Gung Memorial Hospital using routine automated techniques. Hepatitis markers, including IgM class antibody to HAV (IgM anti-HAV) and hepatitis B core antigen (IgM anti-HBc), HBsAg, and antibody to HDV (anti-HD) were assayed using commercial radioimmunoassays (HAVABM, CORABM, Ausria II, Anti-delta, Abbott Laboratories, North Chicago, Illinois). Antibody to HCV (anti-HCV) was assayed using a second-generation enzyme immunoassay (UBI HCV EIA, United Biomedical Inc., New York, New York). We assayed T₄, T₃, FT₄, TSH, and thyroxine binding globulin (TBG) using radioimmunoassays. Thyroid scans and thyroid uptake were studied using Technetium-99m.

Statistical Analysis

Data were expressed with 95% confidence intervals (CIs) or as mean ± SE. The Student t-test, repeated measures analysis of ANOVA, and the Fisher exact test were used to test the statistical significance of observed differences in rates or means.

Results

Top Methods Results Discussion Author & Article Info References

Of the 60 patients with normal baseline AST and ALT levels, 54 had adequate follow-up studies, and of these 54 patients, 15 (28%, CI 16% to 42%) showed ALT elevations 2 months after the start of PTU therapy. None of these patients had clinical symptoms or signs. The peak ALT level in these 15 patients ranged from 0.65 to 3.85 µkat/L (mean, 1.35 ± 0.32 µkat/L). Under close monitoring, the ALT levels of these 15 patients moved toward normal during the following 3 months of propylthiouracil therapy and returned to normal in 13 of the 15 patients (mean, 0.24 ± 0.04 µkat/L) at a reduced dose according to the study protocol (Figure 1). None of these ALT elevations was associated with increased serum bilirubin. Patients with and without ALT elevations showed a similar degree and duration of ALP elevations (Figure 2). Serologic studies showed that none of these episodes was the result of hepatitis A, B, C, or delta virus infection. Autoantibodies indicating autoimmune hepatitis were also not found.

View larger version (37K): [in this window] [in a new window]   Figure 1. Changes of alanine aminotransferase (ALT) levels over time for each of the 15 patients with ALT elevation during propylthiouracil therapy (—-). The solid line with open circles represents the mean level of ALT over time. The solid line with closed circles represents the ALT changes over time in the patient with the highest ALT elevations (231 U/L [3.85 µkat/L]), who received a liver biopsy 5 months after starting therapy.

View larger version (56K): [in this window] [in a new window]   Figure 2. Serial changes of serum alanine aminotransferase (ALT) and alkaline phosphatase (ALP) during propylthiouracil therapy in patients with hyperthyroidism. Lines with open circles indicate 15 patients with abnormal ALT levels at 2 months; closed circles indicate patients without increased ALT. Data are expressed as mean ± SE. The shaded areas represent the normal ranges of serum ALT (< 36 U/L [< 0.60 µkat/L]) and ALP (28 to 94 U/L [0.47 to 1.57 µkat/L]). The overall time comparison of mean ALT levels is significant (P < 0.01 by repeated measures analysis). When the mean ALT levels at different time points are compared, that at month 2 is statistically greater than those at months 1, 3, 4, and 5, and the level at month 3 is also statistically greater than that at month 5 but not than that at month 4 (P < 0.01 by repeated measures analysis).

Liver biopsy was done in only three patients, whose peak ALT levels were 3.85, 0.97, and 0.83 µkat/L, respectively. Biopsy was done when ALT levels were declining (0.87, 0.60, and 0.63 µkat/L, respectively). Histologic examinations showed irregular patches of hemorrhagic necrosis Figure 3 A or ill-defined granulomas Figure 3 B in the perivenular region. The necrotic foci consisted of aggregates of pigmented foamy histocytes and epithelioid cells, with few other inflammatory cell infiltrates. Abundant ceroid and lipofuscin pigments were present in the pigmented foamy histocytes. Cholestasis, endophlebitis, microthrombus, hemosiderin deposition, and tissue eosinophilia were not noted. Mild fatty change was noted in two patients. The portal and periportal regions were unaffected.

View larger version (145K): [in this window] [in a new window]   Figure 3. Histologic findings in patient with hyperthyroidism after propylthiouracil therapy. Panel A. Liver biopsy of the patient shown in Figure 1. Note irregular patches of necrosis in the perivenular region (hematoxylin and eosin, x 200). Panel B. Liver biopsy of another patient showed ill-defined granuloma and mild fatty change (hematoxylin and eosin, x 400). V = central vein.

Discussion

Top Methods Results Discussion Author & Article Info References

The results of our study showed that ALT elevations developed in nearly 30% of patients after 2 months of propylthiouracil administration. Without an untreated control group, which is not possible because of the ethical concerns involved in withholding treatment, one may argue that patients with normal AST and ALT levels at baseline could develop transaminase abnormality due to hyperthyroidism per se [19], as did the patients with ALT elevations before therapy. Such an event is not likely because the mean duration of symptoms before therapy in patients showing ALT elevations during propylthiouracil administration was 2.5 months, 0.2 months longer than in patients with ALT elevations before treatment (data not shown) and 2 months shorter than in patients without ALT elevations after propylthiouracil therapy (see Table 1). If ALT elevations 2 months after propylthiouracil therapy were due to hyperthyroidism per se, the interval between symptom onset and development of ALT elevations would be much greater than that in patients with ALT elevations before therapy. Therefore, the temporal relation suggests that the ALT elevations were induced by propylthiouracil, reflecting subclinical acute hepatocellular injury or at least transaminase abnormality [20]. Although ALP, but not bilirubin, levels also increased in these patients, the liver injury was not considered to be cholestatic [20] because elevation of ALP is a common phenomenon, as shown in patients without ALT elevations (see Figure 2), possibly caused by increased osteoblastic activity after anti-thyroid therapy [21]. The ALP elevations seen in our patients were also due mainly to the increase of bone isoenzyme (Huang MJ. Unpublished data). Results of serologic studies in the patients with ALT elevations after propylthiouracil therapy also confirmed that these changes were not related to hepatitis A, B, C, or delta virus infection. In addition, the six HBsAg-positive patients and two anti-HCV-positive patients all had normal AST and ALT levels during propylthiouracil therapy (Table 1). The latter findings, although not statistically significant, suggest that these ALT elevations were not related to HBV or HCV. The reasons why asymptomatic HBsAg carriers and those who were seropositive for anti-HCV tended to be protected from propylthiouracil hepatotoxicity are unknown. It has been shown that HBsAg carriers have increased cytochrome P-450-dependent function [22], although propylthiouracil may decrease rat cytochrome P-450 and lead to hepatic necrosis [18]. This finding may explain why HBsAg carriers were not affected during propylthiouracil therapy.

Although the incidence of ALT elevations during propylthiouracil therapy was nearly 30%, all patients were asymptomatic and anicteric during the period of ALT elevation. These findings contradict previous observations that propylthiouracil-induced hepatitis was rare and usually severe [1-15]. Perhaps the situation is similar to the observations that mild liver injury manifested by increased serum transaminases levels occurs in approximately 10% to 20%, although clinically apparent hepatitis occurs in fewer than 1% of patients taking isoniazid [5, 23]. Interestingly, evidence has shown that propylthiouracil, like isoniazid, forms active metabolites that may lead to centrilobular hepatic necrosis [18]. The incidence of propylthiouracil-induced mild liver injury appears to be higher than that reported in patients treated with isoniazid; however, hundreds of overt isoniazid hepatitis have been reported [5, 23], whereas only occasional case reports of overt propylthiouracil-induced hepatitis exist [1-15]. By inference, the incidence of clinically apparent hepatitis in patients receiving propylthiouracil might be much lower than that in patients receiving isoniazid. The true incidence of propylthiouracil-induced overt hepatitis remains to be determined by monitoring more patients.

Although liver biopsy was done in only three patients, the findings of perivenular necrosis and granuloma-like lesions are consistent with those of isoniazid-induced hepatitis [5, 23]. The histologic findings observed in our study differ remarkably, however, from the portal hepatitis, periportal hepatitis, or submassive necrosis observed in overt propylthiouracil-induced hepatitis [2-7, 12, 13]. These discrepancies are again consistent with those observed between subclinical and overt isoniazid-induced hepatotoxicity [23].

Previous reports on overt propylthiouracil-induced hepatic injury tended to suggest that propylthiouracil should be discontinued immediately to ensure complete clinical recovery [2-5, 7, 8, 11]. Our results, however, showed that propylthiouracil-induced subclinical ALT elevations normalized gradually during the following 3 months despite continued drug therapy. This result is again similar to that seen in patients with isoniazid-induced mild liver injury [5, 23]. The consistent response pattern of ALT normalization after dose reduction suggests that such propylthiouracil hepatotoxicity is dose related. This finding differs from the reported propylthiouracil-induced overt hepatitis, which has been considered to be an unpredictable allergic host response [4, 6, 14, 16].

Patients with propylthiouracil-induced liver injury had higher pretreatment serum T₃ or T₄ levels than did patients without ALT changes (see Table 1). Hyperthyroidism itself may induce usually mild liver abnormalities, possibly secondary to relative hypoxemia [19]. Patients with higher T₄ or T₃ may have had more severe relative hypoxemia, causing the pericentral area of the liver acinus to become more vulnerable to injury. The histologic findings of focal pericentral necrosis in our patients support this hypothesis.

In conclusion, propylthiouracil-induced mild liver injury is common but, like that of isoniazid, is often asymptomatic, anicteric, and transient despite continued drug administration. Clinically apparent hepatitis was not observed in this series of patients. Regular monitoring of AST and ALT levels has been suggested in patients receiving propylthiouracil to allow discontinuation of treatment at the earliest sign of hepatotoxicity [9, 12]. Our results suggest that propylthiouracil therapy may be continued with caution in patients with asymptomatic and anicteric hepatic injury. Previous reports suggest, however, that treatment should be stopped in patients with overt hepatitis.