1. Top of page
  2. Abstract
  7. References

Objectives To determine reference ranges for liver function tests in uncomplicated pregnancy and to relate abnormal results by these criteria to outcome in pre-eclampsia.

Design Prospective, cross-sectional study to establish the reference ranges. Prospective observational study of women with pre-eclampsia.

Setting Antenatal clinics and obstetric unit of St Mary's Hospital, London.

Participants Four hundred and thirty women with uncomplicated pregnancies and 85 consecutive women with gestational hypertension.

Main outcome measures Aspartate transaminase (AST), alanine transaminase (ALT), bilirubin and gamma glutamyl transferase (GGT) were measured to determine their ranges in normal pregnancy. The severity of pre-eclampsia was determined by the maximum blood pressure, creatinine and 24 h urinary protein; minimum platelet count; maternal complications; mode of and gestation at delivery; and fetal outcome with centile weight adjusted for gestational age and sex.

Results AST, ALT, bilirubin and GGT were each lower in uncomplicated pregnancy than the nonpregnant laboratory reference ranges. Of those cases with elevated liver function tests in the pre-eclampsia group, 37% were abnormal only by the new reference ranges. Using the new ranges, the prevalence of elevated liver function tests was significantly higher in the pre-eclampsia group (54%) than in those with pregnancy induced hypertension (14%) (P < 0.01). Amongst those with pre-eclampsia, abnormal liver function tests were associated with greater proteinuria(P < 0.05), lower platelet count(P < 0.001) and more maternal complications(P < 0.01) than normal liver function tests; there was no difference in the severity of hypertension between the groups.

Conclusions Liver function tests are lower in normal pregnancy than the reference ranges currently used. Our pregnancy-derived ranges allow more precise identification of abnormal liver function in women with pre-eclampsia than is possible using standard reference ranges derived from a nonpregnant population.


  1. Top of page
  2. Abstract
  7. References

Abnormal liver function tests occur in 20% to 30% of pregnancies complicated by pre-eclampsia1,2 and are associated with poor maternal and fetal outcomes2,3. However, there is no consensus as to which levels are abnormal. Weinstein4, in his seminal work on haemolysis, elevated liver enzymes, and low platelets syndrome (HELLP) says only that aspartate transaminase (AST) or alanine transaminase (ALT) or bilirubin were abnormal but does not state their concentrations; Van Dam5 considers AST to be abnormal when above 18 U/L, Goodlin6 uses 30 U/L, Romero2 52 to 57 U/L, and Sibai7 70 U/L. Although these results may not be directly comparable due to the use of different assays with different reference ranges (these facts are not, however, disclosed), the wide spectrum reflects the lack of agreement.

This is largely due to poor definition of the normal ranges despite the fact that this knowledge is fundamental to the correct interpretation of results. In a longitudinal study of 64 women Carter8 states that AST, ALT, gamma glutamyl transferase (GGT) and bilirubin levels do not change during uncomplicated pregnancy from 16 to 40 weeks of gestation and are the same as the nonpregnant values. In a study of 64 women Shukla et al.9 found ALT, and to a lesser extent AST, to be lower in normal pregnancy than in nonpregnant age-matched female controls, and reported that they did not vary with gestation. Cerutti et al.10 published a cross-sectional study of 304 women and concluded that AST, ALT and GGT each show a significant increase from a gestational age of six months but it is not clear whether this is compared with early pregnancy or to the nonpregnant control group. None of these papers define their laboratory reference ranges. Furthermore all have calculated reference ranges using parametric tests on the assumption that the data are normally distributed. On the basis of this conflicting and unconvincing evidence some authorities11 assume that liver function test are not altered by pregnancy and advise using the hospital, nonpregnant, laboratory reference ranges (usually derived from healthy adult men and women of a wide age range).

In the absence of altered hepatic blood flow physiological haemodilution alone may result in lower reference ranges for AST, ALT, GGT and bilirubin in pregnancy if correct statistical methods are used to construct them. If this is so, we wondered whether women whose pre-eclampsia is complicated by even mild abnormalities of liver function test may have more severe disease than women with normal liver function test, or whether previous associations were due to the use of higher ‘normal’ ranges and therefore more severely abnormal results. Therefore we under-took a prospective cross-sectional study to establish normal ranges in pregnancy for AST, ALT, GGT and bilirubin. We used these to investigate prospectively the possibility that in pre-eclampsia even mildly abnormal levels might be associated with poor outcome.


  1. Top of page
  2. Abstract
  7. References

Four hundred and thirty pregnant women gave informed verbal consent for participation in this prospective, cross-sectional study. Women with liver pathology, hypertension or multiple pregnancy were excluded. A single blood sample was taken from each woman at the time of routine venepuncture. AST was measured by the International Federation of Clinical Chemistry (IFCC) recommended method, ALT by the recommended method of the Scandinavian Society of Clinical Chemistry and Clinical Physiology, GGT by the method of Szasz12, and bilirubin by a modi- fication of the method of Jendrassik and Grof13. All assays were performed on a discrete multichannel automated analyser (Olympus AU5200). Alkaline phosphatase was not included because of the estab- lished, but variable, contribution in pregnancy of placental and bone isoenzymes14 and because iso-enzyme assays were not readily available to measure the liver component.

The data for ALT, AST and bilirubin were normally distributed following logarithmic transformation. Gestation-specific 95% reference ranges were calculated for each enzyme, as recommended by the IFCC15, using the modulus exponential normal model. This fits polynomial curves and facilitates the construction of reference ranges for data which are time-dependent16 (AST and bilirubin) or time-independent15,17 (ALT). The minimum number of subjects required to validate this method is 30018. Even after logarithmic transformation the data for GGT were not normally distributed, and the reference range was calculated for each trimester using nonparametric determination of centiles (2.5–97.5)15.

Eighty-five consecutive women presenting with gestational hypertension during the study period were recruited prospectively. Pre-eclampsia was defined as two consecutive measurements of diastolic blood pressure geqslant R: gt-or-equal, slanted 90 mmHg four or more hours apart or a single reading geqslant R: gt-or-equal, slanted 110 mmHg, with proteinuria > 0.3 g per 24 h (or geqslant R: gt-or-equal, slanted 2+ on dipstick testing if a 24 h collection was not completed) in the absence of a urinary tract infection19. All measurements of blood pressure were taken manually by the midwife or doctor caring for the patient (diastolic blood pressure was recorded as Korotkoff four). On each occasion that venepuncture was clinically indicated, AST, ALT, GGT and bilirubin were measured. The new reference ranges were not available to the clinicians and therefore did not influence management.

Outcome variables used were maximum mean arterial blood pressure; maximum creatinine concentration; maximum 24 h urinary protein excretion; minimum platelet count; maternal complications; mode of and gestation at delivery; fetal outcome; birthweight centile adjusted for fetal sex and gestation. The mean arterial blood pressure was calculated as the sum of one third of the difference between the systolic and diastolic pressures, and the diastolic pressure. Maternal complications were considered to be serious medical problems secondary to pre-eclampsia.

The outcome variables of the women with pre-eclampsia, with and without elevated liver enzymes, were compared using χ2 or Fisher's exact test, and Student's t test except for urinary protein and serum creatinine which had a skew distribution and were compared using Wilcoxon's two-sample test. Women with pre-eclampsia and abnormal liver function tests were further divided into two groups: those elevated only by the new criteria and those elevated beyond the traditional laboratory reference ranges.


  1. Top of page
  2. Abstract
  7. References

Four hundred and thirty women each had a single blood sample taken to establish the reference ranges for the liver enzymes ALT, AST, bilirubin and GGT in uncomplicated pregnancy: 120 in the first, 170 in the second and 140 in the third trimester. As a consequence of laboratory technical problems 425 results were available for ALT, 421 for AST, 419 for bilirubin and 393 for GGT. The women represented an unselected sample from our antenatal population. Their average age was 29 years (range 17 to 44 years) and the maximum parity was four. Forty-seven percent were white European, 21% Afro-Caribbean, 10% Mediterranean, 7% Asian, 4% Oriental and 11% other ethnic origin (Fig. 1).


Figure 1. 95% Reference range by gestational age for (a) ALT, (b) AST, (c) bilirubin. In each case, the upper line represents the 97.5th centile, the middle line the 50th and the lower line the 2.5th.

Download figure to PowerPoint

Gestation-specific 95% reference ranges are shown for ALT, AST, and bilirubin in Fig. la, b and c, respectively; their regression equations are show in Table 1. Table 2 shows the nonpregnant ranges used at St Mary's Hospital, the new pregnancy reference range for GGT, and approximate ranges in each trimester for ALT, AST and bilirubin taken from Fig. la, b and c for ease of comparison. Only 10/1658 results were greater than the laboratory nonpregnant normal ranges.

Table 1.  Coefficients for construction of 95% regression equations for ALT, AST and bilirubin: ALT = exp {coeff}; AST = exp {coeff + (0.00256 × gest)}; bilirubin = exp {coeff- (0.040 × gest) + (0.00077 × gest2)}.
Table 2.  Reference ranges for liver enzymes, by trimester.
Liver enzymeNonpregnant1st trimester2nd trimester3rd trimester
ALT (U/L)0-406-326-326-32
Bilirubin (nmol/L)0-174-163-133-14
GOT (U/L)11-505-375-433-41

The prevalence of any abnormal liver function test in the hypertensive women, as defined by the new pregnancy reference ranges, was significantly higher (54%, 19/35) in the women with pre-eclampsia than in those with nonproteinuric hypertension (14%, 7/50) (P < 0.01) and than that previously documented. This is due to our use of pregnancy-specific reference ranges; without these the prevalence in pre-eclampsia was 34% (12/35). Among women with pre-eclampsia those with abnormal liver function tests had significantly greater proteinuria, lower platelet count and more maternal complications than those with normal liver function tests. There was also a trend towards increased neonatal death, caesarean section, and induction of labour; but these differences failed to reach conventional statistical significance; this is possibly because they are binary variables, and with small numbers all but extremely large differences fail to do so. There was no difference in the severity of hypertension between the groups (Table 3). The maternal complications were acute renal failure requiring dialysis(n= 1), profound olig-uria needing central venous pressure monitoring and renal support(n= 4), and spontaneous pulmonary oedema before the use of intravenous fluid(n= 2).

Table 3.  Comparison of outcome variables in women with pre-eclampsia and either normal or abnormal liver function tests (LFT). Values are shown as mean (SD), or median (range) for continuous variables or n (%) for discrete variables. MAP = mean arterial pressure.
Outcome variablesAbnormal LFT (n = 19)Normal LFT (n = 16)
  1. *P < 0.01;**P < 0.001; P < 0.05.

Platelet (x10/L)149** (45)197 (36)
MAP(mmHg)128 (15)124 (10)
Creatinine (umol/L)88 (62-661)85 (53-134)
Proteinuria (g/24 h)2.7 (0.3-10.04)1.3 (0.3-2.63)
Maternal complications7* (37)0
Induction of labour17 (89)11 (69)
Caesarean section12 (63)6 (38)
Gestation at delivery (weeks)35.7 (5.6)37.6 (3.6)
Neonatal death3 (16)0
Weight (centile)32 (22)39 (30)

Seven of the 19 pre-eclamptic women (37%) with abnormal results had values which before this study would have been considered to be normal by our laboratory. They had the same disease spectrum as the women with more grossly elevated levels (Table 4). They still had significantly greater proteinuria (P < 0.05), lower platelet count(P < 0.03) and more maternal complications(P < 0.05) than the group of pre-eclamptic women with normal liver function tests.

Table 4.  Similar spectrum of disease in women with abnormal liver function tests (LFT) defined by old and new normal ranges. Values are shown as mean (SD), or median (range) for continuous variables or n (%) for discrete variables. MAP = mean arterial pressure.
 Abnormal LFT
Outcome variablesBy old range (n = 12)By new range alone (n = 7)
Platelet (xl09/L)148 (42)152 (55)
MAP(mmHg)129 (18)126 (7)
Creatinine (umol/L)81 (62-661)91 (64-121)
Proteinuria (g/24 h)2.7 (0.3-10.04)2.7 (0.6-9.1)
Maternal complications5 (41)2 (29)
Induction of labour10 (83)7 (100)
Caesarean section8 (67)4 (59)
Gestation at delivery (weeks)36.8 (5.3)33.8 (5.8)
Neonatal death2 (17)1 (14)
Weight (centile)34 (23)28 (21)

There was variation between women in the number of elevated results and the order in which they became abnormal. Seventeen women had an abnormal AST, ten ALT, five GGT and six bilirubin. The two women with normal AST had elevated GGT: one case was an isolated finding and the other had raised bilirubin. Overall, seven women had one elevated enzyme, six had two, five had three and one had four.


  1. Top of page
  2. Abstract
  7. References

The accurate definition of the reference range is vital for the correct interpretation of results. This cross-sectional study strongly suggests that in pregnancy liver enzymes are lower than the nonpregnant reference ranges provided by laboratories and used by clinicians as a guide to determining normality.

Unfortunately, neither the means and standard deviations from which the laboratory ranges were derived, nor the demographic details of the population studied, are available for statistical comparison with our data. However, our values are mostly at least 20% lower in normal pregnancy than expected, and less than 1% results exceed the nonpregnant normal ranges. This makes it very likely that the differences we have demonstrated are significant. This may be due to the physiological haemodilution caused by pregnancy or in part to differences in the age and sex distribution of the two populations. This does not obviate the clinically-relevant fact that for healthy pregnant women these lower reference ranges should be used. For serial monitoring of patients at risk reference ranges derived from longitudinal data are required, and these are being constructed at present.

The prevalence of elevated liver function tests in pre-eclampsia in this study is higher than previously documented1,2. Moreover, 20% (7/35) of all women with pre-eclampsia would have the normality of their liver function test results misinterpreted if traditional cutoff points are used. This may have consequences for the management of pre-eclampsia since we have shown that even mildly abnormal liver function tests are associated with worse disease compared with normal values. This is not surprising as it is likely that abnormal liver function reflects vasoconstriction involving the hepatic bed and thus widespread disease. It is also possible that some of the AST and GOT is not of hepatic origin: both are widely distrib- uted throughout the body and may be elevated in relation to pre-eclampsia by haemolysis or endothe- lial damage, respectively20. McMahon21 suggests that in HELLP syndrome early changes in liver function may be due to red cell destruction and that liver damage itself only occurs later. Nonetheless this would not alter the outcome variables associated with abnomal liver function tests. Alternatively, the preva- lence of abnormal liver function tests may be inflated by the use of multiple ranges, which itself increases the probability of classifying as abnormal a woman who is in fact normal: in this case these outcome data would represent an attenuation of the differences between the two groups.

We recommend that pregnancy-specific reference ranges are used for the assessment of liver function in the antepartum period. This may be particularly useful in the management of pre-eclampsia, where underestimation of abnormal liver function will be avoided and more accurate assessment of the severity of the disease possible.


The author would like to thank Professor P. Royston for his expert statistical advice.


  1. Top of page
  2. Abstract
  7. References
  • 1
    Borglin NE. Serum transaminase activity in uncomplicated and complicated pregnancy and in newboms. J Clin Endocrin Metab 1958; 18: 872877.
  • 2
    Romero R, Vizoso J, Emamian M et al. Clinical significance of liver dysfunction in pregnancy induced hypertension. Am J Perinatal 1988; 5: 146151.
  • 3
    Verhaeghe J, Anthony J, Davey DA. Platelet count and liver function tests in proteinuric and chronic hypertension in pregnancy. S Afr Med J 1990; 79: 590594.
  • 4
    Weinstein L. Syndrome of haemolysis, elevated liver enzymes and low platelet count: a severe consequence of hypertension in pregnancy. Am JObstet Gynecol 1982; 142: 159167.
  • 5
    Van Dam PA, Renier R, Baekelandt M, Buytaert P, Jyttenbroek F. Disseminated intravascular coagulation and the syndrome of haemolysis, elevated liver enzymes and low platelets in severe pre-eclampsia. Obstet Gynecol 1989; 73: 97102.
  • 6
    Goodlin RC. Beware the great imitator - severe pre-eclampsia. Contemp Obstet Gynaecol 1982; 20: 215228.
  • 7
    Sibai BH. The HELLP syndrome: much ado about nothing? Am J Obstet Gynecol 1990; 162: 311316.
  • 8
    Carter J. Liver function in normal pregnancy. Aust N Z J Obstet Gynaecol 1990; 30: 296302.
  • 9
    Shukla PK, Sharma D, Mandal RK. Serum lactate dehydrogenase in detecting liver damage in pre-eclampsia. Br J Obstet Gynaecol 1978; 85: 4042.
  • 10
    Cerutti R, Ferrari S, Grella P et al. Behaviour of serum enzymes in pregnancy. Clin Exp Obstet Gynecol 1976; 3: 2224.
  • 11
    Fagan EA. Disorders of the liver and biliary system and pancreas. In deSwietM, editor. Medical Disorders in Obstetric Practice. Oxford : Blackwell Science, 1995: 322.
  • 12
    Szasz G. New substrates for measuring gamma-glutamyl transpepti-dase activity. Z Clin Chem Klin Biochem 1974; 12: 228.
  • 13
    Bartels H, Bohmer M. A simple method for bilirubin determination. Z Klin Chem Klin Biochem 1970; 7: 444447.
  • 14
    Adeniyi FA, Olatunbosun DA. Origins and significance of the increased plasma alkaline phosphatase during normal pregnancy and pre-eclampsia. Br J Obstet Gynaecol 1984; 91: 857862.
  • 15
    International Federation of Clinical Chemistry, Expert Panel on Theory of Reference Values. Part 5: Statistical treatment of collected reference values; determination of reference limits. J Clin Clin Chem 1987; 25: 645656.
  • 16
    Royston P, Wright EM. A parametric method for estimating age-specific reference intervals ('normal ranges'). J Roy Stat Soc (series A) 1997. In press.
  • 17
    Royston P. Parametric models for estimating reference intervals in medicine. Technical Report TR-96–07. London : Imperial College, Statistics Section, Department of Mathematics, 1996.
  • 18
    Royston P. Constructing time-specific reference ranges. Stat Med 1991; 10: 675690.
  • 19
    Davey DA, MacGillivray I. The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 1988; 158: 892898.
  • 20
    Churchill D, Kilby MD, Bignell A et al. Gamma glutamyl transferase activity in gestational hypertension. Br J Obstet Gynaecol 1994; 101: 251253.
  • 21
    McMahon LP, O'Coigligh S, Redman CWG. Hepatic enzymes and the HELLP syndrome: a long-standing error? Br J Obstet Gynaecol 1993; 100: 693695.