A 20-year single-center experience with acute liver failure during pregnancy: Is the prognosis really worse?

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Pregnant patients with acute liver failure (ALF) are believed to have a worse outcome than nonpregnant women and men with ALF. However objective data supporting this supposition are scant. Therefore, the current study compared the outcome, complications, and causes of ALF among pregnant women and girls with age-matched nonpregnant women and girls and men and boys with ALF. One thousand fifteen consecutive ALF patients in the reproductive age group, admitted at the All India Institute of Medical Sciences, New Delhi, from January 1986 to December 2006, were included in the study. A total of 249 (38.5%) women were pregnant. They were compared with 341 nonpregnant women and girls and 425 men and boys, aged 15 to 45 years. The mortality rate of pregnant women and girls (53.8%) was similar to age-matched nonpregnant women and girls (57.2%), and men and boys (57.9%); P = 0.572.The clinical and biochemical features, disease severity, and complications were also similar in the three groups. A significantly higher proportion of ALF was attributable to hepatitis E virus (HEV) among women and girls who were pregnant (59.4%), as compared with both nonpregnant women and girls (30.4%), and men and boys (23.1%); P < 0.001. However, the outcome of HEV-related ALF was independent of the sex and pregnancy status of the patients (P = 0.103). Mortality in HEV-ALF and non–HEV-ALF patients in pregnant women and girls was 51% (74/145) and 54.7% (52/95)(P > 0.1), respectively. The outcome of pregnant ALF patients was also unrelated to the trimester of pregnancy. The mortality of non–HEV-related ALF among the pregnant women and girls (54.7%), age-matched nonpregnant women and girls (61.7%), and men and boys (62.8%) were also similar (P > 0.1). Conclusion: The mortality of pregnant patients with ALF is similar to that of nonpregnant women and girls and men and boys and is independent of the cause or trimester. Pregnancy per se should not be regarded as a poor prognostic factor for a patient with ALF. (HEPATOLOGY 2008.)

Pregnant women with acute liver failure (ALF) are widely regarded as having a more severe disease with higher complication rates and mortality as compared with nonpregnant patients.1, 2 The highest burden of ALF is found in the Indian subcontinent, China, and southeast Asia, where hepatitis E virus (HEV) is the commonest cause of endemic and epidemic acute hepatitis.1, 3–7 Although pregnancy does not confer increased susceptibility to hepatitis A, B, and C viruses, pregnant women may be more vulnerable to infection by HEV. During epidemics of acute non-A, non-B viral hepatitis, a high attack rate and mortality has been noted during pregnancy.3–6 Most of these epidemics were presumably caused by HEV infection. Similar findings have also been noted in the sporadic setting, but not consistently.7, 8 Furthermore, HEV hepatitis is a more severe disease during pregnancy.3–6 Pregnant women and girls with acute HEV hepatitis decompensate to liver failure more commonly than nonpregnant patients with acute hepatitis.3–6

However, progression from exposure to infection, development of clinical acute hepatitis, development of liver failure, and eventual recovery or demise involves a series of differential susceptibilities of the host to the hepatic insult. Increased vulnerability to infection should not be extrapolated to imply a reduced chance of recovery from liver failure. The outcome of the patient may no longer depend on the cause and pregnancy status once liver failure develops. No study has specifically described the prognosis of pregnant women and girls who develop ALF. This issue is crucial, because if pregnancy per se confers a poor prognosis, then the threshold for liver transplantation for pregnant ALF patients may be lower. Second, termination of pregnancy may be indicated, if pregnancy confers a higher mortality risk. To answer these questions, we need a direct comparison of the outcome of consecutive pregnant women and girls, nonpregnant women and girls, and men and boys with ALF. Such a study design can also decide whether the course of ALF attributable to a particular cause is different during pregnancy.

The aim of the current study was to compare the causes, complications, and outcome of consecutive pregnant patients with ALF, with nonpregnant women and girls and men and boys with ALF, in a sporadic setting.

Abbreviations

ALF, acute liver failure; HAV, hepatitis A virus; HEV, hepatitis E virus; IgM, immunoglobulin M; SD, standard deviation.

Patients and Methods

From January 1986 to December 2006, 1155 patients with ALF were admitted in the Department of Gastroenterology, All India Institute of Medical Sciences, New Delhi. Age between 15 and 45 years was considered as the reproductive period. To compare the outcome of pregnant women and girls with nonpregnant patients of a comparative age-group, we included a total of 1015 consecutive patients with ALF between 15 and 45 years of age in this study. The excluded 140 patients were either younger than 15 years (n = 26), or older than 45 years of age (n = 114).

Study Variables.

ALF was defined by the occurrence of encephalopathy within 4 weeks of onset of symptoms in the absence of preexisting liver disease.9 The diagnosis was confirmed by the presence of submassive or massive necrosis in the postmortem liver biopsy specimens of patients who died.

Pre-encephalopathy and icterus-encephalopathy periods were defined as the interval from the onset of prodrome and jaundice, respectively, to the onset of hepatic encephalopathy.7

Admission to death interval was defined as the time from admission to death among the nonsurvivors.

Consciousness recovery time was defined as the period between admission and improvement to grade 0 encephalopathy.

Grading of encephalopathy was done as follows10–12:

  • Grade 1: Loss of sleep rhythm, drowsiness, confusion, and flapping tremors

  • Grade 2: Features of grade 1 encephalopathy with loss of sphincter control

  • Grade 3: Unconsciousness with no response to oral commands, but responding to painful stimuli

  • Grade 4: Deep unconscious state, with no response to pain

Cerebral edema was defined clinically by the presence of spontaneous or inducible decerebrate posturing, or by the presence of any two of: hypertension (blood pressure ≥150/90 mmHg), bradycardia, pupillary changes, or neurogenic hyperventilation.7

Infection was diagnosed by the presence of pyrexia (temperature > 101°F) or hypothermia (temperature < 98°F) and neutrophil leukocytosis (total leukocyte count >15,000/mm3, with ≥ 80% polymorphs), and one or more of: positive blood culture, positive urine culture, or radiological evidence of pneumonia.7, 12

Renal failure was diagnosed if patients developed decreased urine output (< 400 mL in 24 hours), with serum creatinine greater than 1.5 mg/dL and blood urea greater than 40 mg/dL, despite hydration, objectively assessed by central venous pressure of 10 cm of saline or more.11, 12

Management Protocol.

All patients were admitted to the gastroenterology intensive care unit, and a uniform management protocol was followed. History was obtained from the attendants of the patient, and a detailed clinical examination was performed at admission. All patients were subsequently evaluated every 2 hours. Each patient had continuous, noninvasive cardiac, oxygen saturation, and blood pressure monitoring. Blood glucose was monitored every 2 hours. Neurological assessment was done once daily and on any sign of deterioration. The parameters noted were: grade of encephalopathy, presence of spontaneous or induced decerebration, pupillary size and reaction, and presence of any focal neurological deficit. Complete blood counts, biochemistry, and body fluid surveillance cultures were obtained daily.

Elective ventilation was used for all patients in grade 4 encephalopathy and those in grade 3 encephalopathy with evidence of cerebral edema. Cerebral edema was managed conservatively by mannitol given in standard doses. Lactulose or other ammonia-lowering therapies were not used in any case. Liver transplantation and liver replacement therapies were unavailable, and all patients were followed up until recovery or death.

Pregnancy was diagnosed by documentation of fetus in the uterus by bedside ultrasonography, and the gestational period was calculated from the duration of amenorrhea. The duration of amenorrhea was obtained from the near relatives of the patient.

Pregnancy was managed conservatively, without active induction of labor. If the membranes ruptured or the patient developed signs of amnionitis, then labor was induced by oxytocin infusion or prostaglandins. During active labor, an effort was made to correct coagulation parameters by fresh frozen plasma infusions.

Viral Markers.

A serum sample from each patient was tested for hepatitis B surface antigen, immunoglobulin M (IgM) antibody against hepatitis B core antigen, and IgM antibody against hepatitis A virus (IgM anti-HAV) using commercial immunosorbent assay (enzyme-linked immunosorbent assay) test kits (Organon, Teknika, Netherlands). IgM antibody to ORF-1, ORF-2, and ORF-3 of HEV and HEV RNA were tested by methods developed at our institute.13, 14 Anti-hepatitis C antibody was tested by using a third-generation, commercial enzyme-linked immunosorbent assay (Xcyton, Bangalore, India) method. Hepatitis C virus RNA was tested by reverse transcription nested polymerase chain reaction, using primers from the 5′ nontranslated region, if appropriate by methods already described.7

Appropriate tests for autoimmune hepatitis (anti-nuclear factor, anti-smooth muscle antibody, anti-liver kidney microsomal antibody), and Wilson's disease (serum ceruloplasmin, 24-hour urinary copper and serum copper) were also performed in each case. Each patient had bedside ultrasonography.

Data Analysis and Statistics.

Normal distribution of data was confirmed by the Kolmogorov-Smirnov test. The primary objective of interest was survival of pregnant ALF patients. For comparison we used age-matched (15-45 years) nonpregnant women and girls, and men and boys. The proportions of survivors in the three groups were compared using the chi-squared test. Homogeneity of variance or homoscedasticity was tested by Levene's test. Numerical or ordinal data were compared between groups by the one-way analysis of variance. For nonhomoscedastic data, the Kruskal-Wallis test was used for comparison. Specific pairwise post-hoc comparisons were performed using the Bonferroni test. All P-values are two-sided, with a significance level of 0.05.

Statistical software SPSS (version 11.5; SPSS Inc., Chicago, IL), was used for statistical analysis.

Results

During the study period, 1015 ALF patients, aged 15 to 45 years, were admitted at our unit. These included 590 women or girls (58.1%), of whom 249 (42.2%) were pregnant. The baseline clinical features and disease causes of the ALF patients are summarized in Tables 1 and 2.

Table 1. Baseline Clinical Features of Patients with ALF (n = 1015)
  • *

    Etiological evaluation available for 993 patients.

  • Ten patients IgM HAV and IgM HEV positive; twenty-nine patients IgM HBc and IgM HEV positive; seven patients IgM HBc, IgM HEV, and anti-HCV positive; two patients IgM HAV and IgM HBc positive; one patient IgM HAV, IgM HBc, and IgM HEV positive.

  • Ninety-four patients only HBsAg positive; two patients only anti-HCV positive; three patients HBsAg and anti-HCV positive.

Age (years) 
 Mean25.6
 SD6.7
Sex (Females)590 (58.1%)
Pregnancy249/590 (42.2%)
Pre-encephalopathy interval (days) 
 Mean7.7
 SD6.3
Icterus-encephalopathy interval (days) 
 Mean6.1
 SD5.9
Grade 3,4 encephalopathy (at admission)810 (79.8%)
Cerebral edema present (at admission)483 (47.6 %)
Serum bilirubin (mg/dL) 
 Mean15.9
 SD8.9
Serum alanine aminotransferase (IU/L) 
 Mean956.5
 SD110.2
Serum albumin (mg/dL) 
 Mean3.0
 SD0.7
Blood urea (mg/dL) 
 Mean28.1
 SD25.3
Serum creatinine (mg/dL) 
 Mean1.0
 SD1.1
Serum sodium (meq/L) 
 Mean139.5
 SD7.5
Etiology* 
 Acute A12 (1.2%)
 Acute B48 (4.8%)
 Acute E342 (34.4%)
 Dual acute49 (4.9%)
 Only chronic markers99 (10.0%)
 Drugs/toxins44 (4.4%)
 Etiological evaluation negative399 (40.2%)
Table 2. Outcome of HEV and Non-HEV Patients with ALF in Age-Matched Pregnant, Nonpregnant Women and Girls, and Men and Boys
 Men and BoysNonpregnant Women and GirlsPregnant Women and GirlsTotal
  • *

    Frequency of HEV-related ALF significantly higher among pregnant women and girls as compared with nonpregnant women and girls (P < 0.001).

  • Frequency of HEV-related ALF significantly higher among pregnant females as compared with men and boys (P < 0.001).

  • Mortality of pregnant women and girls with HEV-related ALF similar to mortality of pregnant women and girls with non−HEV-related ALF (P = 0.5743; odds ratio for mortality: 0.86, confidence interval: 0.50-1.5).

  • §

    Mortality among men and boys, nonpregnant women and girls, and pregnant women and girls with HEV-ALF was similar (P = 0.103).

  • Mortality among men and boys, nonpregnant women and girls, and pregnant women and girls with non-HEV ALF was similar (P = 0.367).

Total5093972491155
Age 15-45 years (reproductive age group)4253412491015
Complete etiological evaluation420 (98.8%)329 (96.5%)244 (98.0%)993 (97.8%)
HEV-related ALF97/420 (23.1%)100/329 (30.4%)145/244 (59.4%)*342/993 (34.4%)
Non-HEV ALF293/420 (69.8%)214/329 (65.0%)95/244 (38.9%)602/993 (60.6%)
Dual acute infections30/420 (7.1%)15/329 (4.6%)4/244 (1.6%)49/993 (4.9%)
Mortality of HEV-related ALF§36/97 (37.1%)46/100 (46.0%)74/145(51.0%)156/342 (45.6%)
Mortality of non-HEV ALF184/293 (62.8%)132/214 (61.7%)52/95 (54.7%)368/602 (61.1%)

Overall, 575 (56.7%) patients died, with mean admission to death interval of 5.1 [standard deviation (SD), 4.2] days. The complications of infection occurred in 309 (30.4%), gastrointestinal bleeding in 73 (7.2%), and seizures in 118 (11.6%) patients. The most common cause of ALF during pregnancy was HEV (Tables 2 and 3). A significantly higher proportion of ALF was caused by HEV among the pregnant women and girls (59.4%, 145/244), as compared with both age-matched nonpregnant women and girls (30.4%, 100/329), and men and boys (23.1%, 97/420); P < 0.001(Table 2).

Table 3. Comparison Demography, Liver Function Tests, and Cause of Disease of Pregnant Women and Girls with Age-Matched Nonpregnant Women and Girls and Men and Boys (n = 1015)
 Men and Boys (n = 425)Nonpregnant Women and Girls (n = 341)Pregnant Women and Girls (n = 249)Significance
  • *

    Using Kruskall-Wallis test.

  • Using ANOVA.

  • Etiological evaluation available for 993 patients.

  • §

    Two or more of IgM-HAV, IgM-HEV, and IgM-HBc positive.

  • Only HBsAg positive or anti-HCV positive.

Pre-encephalopathy interval (days)    
 Mean7.97.87.20.309*
 SD6.46.75.8 
Icterus to encephalopathy interval (days)    
 Mean6.06.95.20.071*
 SD6.06.64.4 
Admission to death time (days)    
 Mean4.95.15.40.574
 SD3.94.64.2 
Consciousness recovery time (days)    
 Mean2.12.02.50.124
 SD2.41.82.4 
Serum bilirubin (mg/dL)    
 Mean16.915.914.40.149*
 SD10.58.45.8 
Serum alanine aminotransferase (IU/L)    
 Mean1268.6819.3613.1<0.001*
 SD1323.3991.9625.7 
Serum alkaline phosphatase (U/L)    
 Mean187.9212.8257.6<0.001
 SD179.4214.3213.1 
Serum albumin (mg/dL)    
 Mean3.22.92.7<0.001
 SD0.60.60.6 
Blood urea (mg/dL)    
 Mean29.830.621.9<0.001*
 SD25.829.315.9 
Serum creatinine(mg/dL)    
 Mean1.11.00.90.096
 SD1.21.20.6 
Serum sodium    
 Mean139.1140.1139.30.586
 SD6.88.18.0 
Etiology    
 Acute A4 (1.0%)6 (1.8%)2 (0.8%) 
 Acute B21 (5.0%)20 (6.1%)7 (2.9%) 
 Acute E97 (23.1%)100 (30.4%)145 (59.4%) 
 Dual acute§30 (7.1%)15 (4.6%)4 (1.6%) 
 Only chronic markers52 (12.4%)35 (10.6%)12 (4.9%) 
Drugs/toxins10 (2.4%)28 (8.5%)6 (2.5%) 
Etiological evaluation negative206 (49.0%)125 (38.0%)68 (27.9%) 

Comparative Mortality of Pregnant Women and Girls With ALF.

Of all the patients in the reproductive age-group, 134 of 249 (53.8%) pregnant women and girls, 195 of 341 (57.2%) age-matched nonpregnant women and girls, and 246 of 425 (57.9%) age-matched men and boys died (Table 4, P = 0.572). Pregnancy status did not confer increased mortality risk for women and girls (odds ratio for mortality, 0.87; 95% confidence interval, 0.62-1.23).

Table 4. Disease Severity and Complications Among Pregnant Women and Girls and Age-Matched Nonpregnant Women and Girls and Men and Boys (n = 1015)
 Men and Boys (n = 425)Nonpregnant Women and Girls (n = 341)Pregnant Women and Girls (n = 249)Significance
Cerebral edema at admission198 (46.6%)164 (48.1%)121 (48.6%)0.819
Encephalopathy at admission    
 Grades 1,295 (22.4%)63 (18.5%)47 (18.9%)0.346
 Grades 3,4330 (77.6%)278 (81.5%)202 (81.1%) 
Seizures54 (12.7%)37 (10.9%)27 (10.8%)0.660
Gastrointestinal bleed38 (8.9%)20 (5.9%)15 (6.0%)0.187
Renal failure52 (12.2%)21 (6.2%)21 (8.4%)0.014
Infection114 (26.8%)107 (31.4%)88 (35.3%)0.061
Mortality246 (57.9%)195 (57.2%)134 (53.8%)0.572

Comparison of Disease Severity and Outcome.

The clinical and biochemical features of pregnant women and girls with ALF are compared with age-matched men and boys, and nonpregnant women and girls with ALF in Table 3. There was no difference in the rapidity of onset of encephalopathy, frequency of cerebral edema, and grades of encephalopathy at presentation among the three groups. Frequency of disease complications such as infection, seizures, and gastrointestinal bleed were also similar among the three groups (Table 4).

Outcome of HEV-Related and Non–HEV-Related ALF During Pregnancy.

A causative evaluation was available for 993 (97.8%) of the 1015 study patients (Table 2). The outcome of pregnant women and girls with ALF was unrelated to the cause of disease. Excluding four pregnant patients with dual acute infections, the mortality rate of 145 HEV-related and 95 non–HEV-related pregnant women and girls with ALF were 51.0% and 54.7%, respectively (P = 0.5743; odds ratio for mortality, 0.86; 95% confidence interval, 0.50-1.50).

Outcome of HEV-Related ALF and Pregnancy Status.

There were 342 patients with HEV-related ALF in the age group of 15 to 45 years (Table 2). The mortality rate of HEV-related ALF was similar among pregnant women ands girls (74/145, 51.0%), nonpregnant women and girls (46/100, 46.0%), and men and boys (36/97, 37.1%) (P = 0.103).

Outcome of Non–HEV-Related ALF and Pregnancy Status.

There were 602 patients with ALF attributable to causes other than HEV (Table 2). The mortality rate of non–HEV-related ALF was similar among pregnant women and girls (52/95, 54.7%), nonpregnant women and girls (132/214, 61.7%), and men and boys (184/293, 62.8%) (P = 0.367).

Cause of Disease and Outcome Related to Trimester.

Of the 249 pregnant women and girls, the mortality rate was similar in the first (3/5, 60.0%), second (92/171, 53.8%), and third (39/70, 55.7%) trimesters (P = 0.619). The mean gestational periods of the 115 pregnant survivors and 134 pregnant nonsurvivors were 26.2 (SD 6.3) weeks and 27.1 (SD 6.0) weeks, respectively (95% confidence interval of mean difference, 0.7 to 2.4; P = 0.289) (Table 5).

Table 5. Comparisons of Survivors and Nonsurvivors Among Women and Girls (n = 590)
ParameterSurvivors (n = 261)Nonsurvivors (n = 329)Significance
  • *

    For 249 of 590 females who were pregnant.

  • Five hundred fifty-four patients, after excluding patients with dual acute infections.

Age (years)   
 Mean25.526.30.089
 SD5.56.2 
Pregnancy115 (44.1%)134 (40.7%)0.416
Gestational period*   
 Mean26.227.10.289
 SD6.36.0 
Cerebral edema at admission87 (33.3%)199 (60.5%)<0.001
Encephalopathy at admission   
 Grades 1,274 (28.4%)36 (10.9%)<0.001
 Grades 3,4187 (71.6%)293 (89.1%) 
Pre-encephalopathy interval (days)   
 Mean7.17.80.171
 SD6.06.5 
Icterus to encephalopathy interval (days)   
 Mean5.56.70.014
 SD4.96.4 
Serum bilirubin (mg/dL)   
 Mean13.416.7<0.001
 SD6.97.6 
Serum alanine aminotransferase (IU/L)   
 Mean746.5719.50.722
 SD879.0840.7 
Serum alkaline phosphatase (U/L)   
 Mean227.2236.90.0625
 SD195.2230.4 
Serum albumin (mg/dL)   
 Mean2.82.80.595
 SD0.60.6 
Blood urea (mg/dL)   
 Mean27.426.50.692
 SD24.724.9 
Serum creatinine (mg/dL)   
 Mean0.91.00.044
 SD0.71.2 
Serum sodium   
 Mean139.9139.40.662
 SD7.78.3 
Prothrombin time prolongation (seconds)   
 <310<0.001
 3-255415 
 25-602627 
 >601644 
Etiology   
 Non-HEV125 (50%)184 (60.5%)0.013
 HEV125 (50%)120 (39.5%) 

Comparison of Survivors and Nonsurvivors Among Women and Girls.

The factors that affect mortality are listed in Table 5. Presence of cerebral edema and grade of encephalopathy at admission, icterus-to-encephalopathy interval, bilirubin and creatinine levels, prothrombin time prolongation, and HEV versus non-HEV etiology significantly influenced outcome among women and girls. Women and girls with higher grades of encephalopathy (grades 3 and 4), presence of cerebral edema at admission, longer icterus to encephalopathy interval, higher serum bilirubin, creatinine, and more prolonged prothrombin time were associated with significantly higher mortality (Table 5). Pregnancy did not confer increased mortality risk among women and girls (Table 5).

Discussion

The current single-center study of a large number of predominantly hepatitis virus–associated ALF patients has shown that the mortality of pregnant women and girls with ALF is similar to the mortality of age-matched nonpregnant women and girls and men and boys with ALF. Women and girls who were pregnant were younger than the nonpregnant women and girls [24.9 (SD 12.0) years versus 29.6 (SD4.5) years; P < 0.001]. Therefore, we included only age-matched nonpregnant women and girls for comparison. Comparison with men and boys admitted during the same time period showed that sex did not influence outcome. No previous study has found sex to be an important independent prognostic variable.

The cause of liver failure differs across the globe accordingly to the prevalent hepatotropic virus infections and drug usage pattern. In the Indian subcontinent and large areas of the developing world, HEV is the most common cause of endemic and epidemic acute hepatitis. HEV is a waterborne infection, and everyone would be equally exposed regardless of their sex and pregnancy status. However, we found that HEV was the causative agent of ALF among approximately 60% of pregnant women and girls, in contrast to 23% and 30% among age-matched men and boys and nonpregnant women and girls, respectively (Table 2). This high proportion of HEV-related ALF among pregnant patients may be attributable to a higher attack rate of HEV and a greater progression of HEV-related hepatitis to liver insufficiency during pregnancy. The higher predisposition of pregnant women and girls to severe sequelae of hepatic insults also leads to a disproportionate pregnancy rate among women and girls with ALF. Of the 590 women and girls with ALF in the age-group of 15 to 45 years, 249 (42.2%) were pregnant (Table 2). In an earlier series from India, 60% women of childbearing age with ALF were pregnant.15 In contrast, the fertility rate among women of child-bearing age in the general population of India is 2.9%.16 HEV is predominantly transmitted through contaminated water and food.1–7, 14, 15, 17–21 The disproportionate pregnancy rate among the women and girls with ALF in the current study may raise the suspicion of sexual transmission of HEV. However, sexual transmission of HEV has not been reported. HEV viremia persists for only approximately 2 weeks after the onset of hepatic illness, which is likely not long enough for sexual transmission of the infection.

Most authorities believe that HEV causes a more severe disease during pregnancy.17–22 However, we found that although HEV-related ALF is more common during pregnancy, once the liver fails the subsequent prognosis is independent of the cause. Both HEV-related and non–HEV-related ALF during pregnancy had similar disease severity at presentation, frequency of complications, and mortality rates. The mortality in pregnant patients with HEV-related ALF was 51%, which was similar to the 54.7% mortality rate of pregnant patients with non–HEV-related ALF (P > 0.1) (Table 2). Hence, the cause of liver failure was unrelated to patient outcome during pregnancy. In contrast, Khuroo and Kamili15 found that the survival of all patients with HEV-related ALF was better than ALF attributable to other causes. However, their study only included 49 pregnant women and girls, of whom two patients had non-HEV causes.15

The previously held belief that viral hepatitis has a worse outcome during pregnancy cannot be discounted or supported by the current study. The course of viral hepatitis in pregnancy has three components. The first issue is whether pregnant women are more susceptible to viral hepatitis. Our study was not designed to answer this question. Sporadic acute viral hepatitis caused by agents other than HEV does not show any special predilection for pregnant women and girls. However, a higher attack rate among pregnant women and girls, as compared with nonpregnant women and girls and men and boys during epidemics of HEV, and possibly also during endemic conditions, has been reported from India and elsewhere.1–6, 23, 24 Khuroo and Kamili15 reported that the cause of sporadic acute hepatitis was HEV among 85.5% pregnant women and 41.5% nonpregnant women in Kashmir, India. Conversely, pregnant women are less likely to be exposed to hepatotoxic drugs.

Second, do pregnant women with viral hepatitis more often progress to liver failure? There are a few studies in the literature to support this notion, but all of them suffer from selection bias. Most studies describing the course of acute viral hepatitis during pregnancy have included self-referred patients. Patients who have more severe illness are more likely to report to the hospital and progress to liver failure. However, there is some evidence that acute hepatitis caused by HEV may be particularly severe among pregnant women and girls, with mortality rates reaching 15% to 20%.25 In prospective studies evaluating pregnant patients with acute hepatitis caused by HEV, progression to ALF has been reported in 15% to 60% of patients.26–29 In two prospective, comparative studies from Kashmir and Delhi, 55% to 70% of pregnant women and girls with HEV hepatitis compared with 10% to 20% of pregnant women and girls with non-HEV hepatitis progressed to liver failure.15, 29 Hence, increased attack rates of HEV hepatitis and increased progression to liver failure during pregnancy together translate into a higher proportion of ALF during pregnancy in endemic areas. A shift in T-helper cell response toward Th2 response, diminished cellular immunity as indicated by a decrease in CD4 cells, increase in CD8 cells, and lowered CD4/CD8 ratio among pregnant women and girls with HEV-related ALF compared with HEV-negative pregnant ALF patients have been reported.30, 31 It is also possible that in utero fetal transmission of HEV with fetal hepatitis may aggravate the maternal hepatitis.32

The current study addressed the third issue of whether liver failure during pregnancy had a worse prognosis. The disproportionate mortality attributable to ALF among pregnant women and girls in endemic areas is because of their high absolute numbers. HEV leads to a large number of deaths among pregnant ALF patients simply because it is the major cause of ALF in endemic areas. We found that the severity of liver failure and complications were similar among the pregnant women and girls, age-matched nonpregnant women and girls, and men and boys regardless of the cause (Table 4). The mortality rate of HEV-related ALF was similar among pregnant women and girls, nonpregnant women and girls, and men and boys (Table 2). This fact highlights the fact that the influence of pregnancy becomes redundant once liver failure develops. In contrast, Jilani et al.31 recently reported that HEV-related pregnant ALF patients had a higher mortality in comparison with HEV-related nonpregnant women and girls. However, their study only included 38 pregnant women and girls and 15 nonpregnant women and girls with HEV-related ALF.

Most of the liver failures in India are caused by hepatitis viruses, and we did not find acute fatty liver of pregnancy as a cause of ALF among our pregnant patients, which has been reported earlier.7, 11, 12, 15 Acute fatty liver of pregnancy has a strong genetic component, and may have geographical variation in prevalence.33 There is thus no rationale for actively terminating pregnancy in HEV endemic regions with the hope of improving the outcome of the patient. Inducing labor may expose the patient to the risk of severe postpartum hemorrhage, because of the attendant coagulopathy. An attempt to correct the coagulopathy inevitably leads to a large volume burden and may exacerbate intracranial hypertension. Valsalva straining during expulsion of fetus and instrumental manipulation may further raise the intracranial pressure. It is unlikely that randomized trials can be ethically conducted comparing actively terminating pregnancy with expectant pregnancy management on the survival of patients with ALF.

We found a very low frequency of HAV-related ALF among pregnant female patients. We have earlier shown a high rate of anti-HAV antibodies among children in north India, attributable to near universal childhood infection.34 Similar experience has been reported from other parts of India. Recently universal testing of female patients before pregnancy and vaccination of nonimmune women and girls has been recommended.35 This strategy may be unnecessary for India. Similarly, a low rate of hepatitis B virus–related ALF during pregnancy may be attributable to lower percutaneous interventions during pregnancy and preexisting immunity.

There was no relation of duration of gestation with the maternal outcome in pregnant ALF patients overall, or among those with HEV-related disease. This is in contrast to earlier reports in which a higher maternal mortality was reported in the third trimester as compared with the first trimester.

It is of great public health importance that the large number of deaths among pregnant women and girls that are caused by ALF can be eliminated simply by ensuring clean drinking water supply. The latest data from the National Sample Survey Organization for 1998 indicates that only 70% of urban and 18.7% of rural households have access to piped water supply in India.

In the current study, approximately 40% of all patients with ALF did not have serological markers of hepatitis A through E (Table 1). None of these patients with indeterminate cause of disease had a history of consumption of any indigenous formulations or known hepatotoxic agents. A history of any definite parenteral exposure was also absent in them. All of the seronegative patients presented with clinical features suggestive of acute viral hepatitis (that is, prodrome followed by icterus). They could have been caused by an as-yet-unidentified non-A, non-E hepatitis virus transmitted feco-orally. It is also possible that a small proportion of them could have been caused by HEV itself. During established epidemics of HEV, approximately 15% of icteric patients have been reported to be IgM-HEV negative.36

The current study also showed certain distinct demographic, causative, and clinical characteristics that are different from those reported in patients with ALF from the United States and Europe. In a large multicenter study that included 308 ALF patients, reported by the acute liver failure study group in the United States, the mean age of the patients was 38 years, in contrast to the mean age of 25.6 years in the current study. In the current study, 80% of the patients had advanced encephalopathy (grade 3 or 4), and 48% had cerebral edema at presentation (Table 1), whereas in the above-mentioned US multicenter study, 48% of the patients presented with advanced encephalopathy, and approximately 25% had cerebral edema at admission. This difference indicates that our patients are probably presenting late, which may influence the outcome, as has been reported earlier.11, 37

The causes of ALF in India and the west are also distinct. In India, viral hepatitis is the cause of ALF in approximately 95% of the patients, as opposed to the west, where the cause of ALF is heterogenous.38–41 In the west, hepatitis virus have been implicated in only 4% to 36% of patients.39, 41 In a recent report from the US acute liver failure study group including 1147 adult patients, 46% were caused by paracetamol overdose, 11% were caused by idiosyncratic drug reaction, and only 10% were caused by hepatitis virus (HAV and hepatitis B virus) infection.39 In the United Kingdom, approximately 70% of the ALF was attributable to paracetamol.41 Autoimmune hepatitis and Wilson's disease constitute approximately 7% to 10% of ALF in the west, whereas in India such a cause is infrequent.40 HEV as a cause of ALF has not been reported from the west, whereas it is the commonest cause of ALF in the current study as well as earlier reported studies from India.3–7, 11, 12, 14, 15 At our center, liver transplant facilities are not available; therefore, the mortality rate of 57% documented in the current study would indicate the true frequency of death attributable to ALF. In the US multicenter study, 25% to 29% of patients with ALF were subjected to liver transplantation.38, 39 Forty-three percent of the ALF patients managed with supportive therapy and 70% of the transplanted patients survived.38, 39

The strengths of the current study include the large number of patients studied. Our study including 249 consecutive pregnant patients is the largest reported so far. The groups for comparison were similarly managed in the same center. Liver transplantation was not available, and all patients were followed until recovery or death. Our results can be generalized to most of the patients with postviral ALF in developing countries, which have the highest burden of this disease along with the least resources.

In conclusion, pregnancy does not influence the severity and outcome of ALF and should not be considered as a poor prognostic variable. The management of pregnant patients with ALF should be similar to that of nonpregnant patients.

Ancillary