Plasma uric acid remains a marker of poor outcome in hypertensive pregnancy: a retrospective cohort study

Authors

  • TL-A Hawkins,

    1. Departments of Medicine and Obstetrics and Gynaecology, University of Calgary, Alberta, Canada
    2. Departments of Renal Medicine and Medicine, St George Hospital and University of New South Wales, Kogarah, NSW, Australia
    Search for more papers by this author
  • JM Roberts,

    1. Department of Obstetrics, Gynaecology and Reproductive Sciences, and Epidemiology, Magee-Women’s Research Institute, University of Pittsburgh, Pittsburgh, PA, USA
    Search for more papers by this author
  • GJ Mangos,

    1. Departments of Renal Medicine and Medicine, St George Hospital and University of New South Wales, Kogarah, NSW, Australia
    2. Department of Women’s and Children’s Health, St George Hospital, Kogarah, NSW, Australia
    Search for more papers by this author
  • GK Davis,

    1. Department of Women’s and Children’s Health, St George Hospital, Kogarah, NSW, Australia
    Search for more papers by this author
  • LM Roberts,

    1. Department of Women’s and Children’s Health, St George Hospital, Kogarah, NSW, Australia
    Search for more papers by this author
  • MA Brown

    1. Departments of Renal Medicine and Medicine, St George Hospital and University of New South Wales, Kogarah, NSW, Australia
    2. Department of Women’s and Children’s Health, St George Hospital, Kogarah, NSW, Australia
    Search for more papers by this author

Dr MA Brown, Department of Renal Medicine, St George Hospital, Kogarah, NSW 2217, Australia. Email mbrown@unsw.edu.au

Abstract

Please cite this paper as: Hawkins T, Roberts J, Mangos G, Davis G, Roberts L, Brown M. Plasma uric acid remains a marker of poor outcome in hypertensive pregnancy: a retrospective cohort study. BJOG 2012 2012;119:484–492.

Objective  To examine the relationship between hyperuricaemia, haemoconcentration and maternal and fetal outcomes in hypertensive pregnancies.

Design  Retrospective analysis of a database of hypertensive pregnancies.

Setting  St George Hospital, a major obstetric unit in Australia.

Population  A cohort of 1880 pregnant women without underlying hypertension or renal disease, referred for management of pre-eclampsia or gestational hypertension.

Methods  Demographic, clinical and biochemical data at time of referral and delivery were collected for each pregnancy. Women were grouped according to diagnosis (pre-eclampsia or gestational hypertension) and logistic regression analysis was used to determine the relationship between uric acid, haemoglobin, haematocrit and adverse outcomes; an α level of P < 0.01 was used for statistical significance.

Main outcome measures  Composites of adverse maternal and fetal outcomes.

Results  In women with ‘benign’ GH (without proteinuria or any other maternal clinical feature of pre-eclampsia) gestation-corrected hyperuricaemia was associated with increased risk of a small-for-gestational-age infant (OR 2.5; 95% CI 1.3–4.8) and prematurity (OR 3.2; 95% CI 1.4–7.2), but not with adverse maternal outcome. In the whole cohort of hypertensive pregnant women (those with pre-eclampsia or gestational hypertension) the risk of adverse maternal outcome (OR 2.0; 95% CI 1.6–2.4) and adverse fetal outcome (OR 1.8; 95% CI 1.5–2.1) increased with increasing concentration of uric acid. Hyperuricaemia corrected for gestation provided additional strength to these associations. Haemoglobin and haematocrit were not associated with adverse pregnancy outcome.

Conclusions  Hyperuricaemia in hypertensive pregnancy remains an important finding because it identifies women at increased risk of adverse maternal and particularly fetal outcome; the latter, even in women with gestational hypertension without any other feature of pre-eclampsia.

Introduction

The hypertensive disorders of pregnancy, which include gestational hypertension and pre-eclampsia, increase obstetric risk. Pre-eclampsia is a leading cause of maternal and perinatal morbidity and mortality, and occurs in 2–8% of all pregnancies.1 Hyperuricaemia has been described commonly in pre-eclamptic pregnancies, often preceding the diagnosis of pre-eclampsia and historically was used as a diagnostic marker of pre-eclampsia. The diagnostic role of uric acid in pre-eclampsia diminished because it was thought to be a less important marker of maternal hypertensive renal injury than was proteinuria. The concept that hyperuricaemia may not just be a marker of the pre-eclampsia syndrome, but may contribute to its pathogenesis was recently introduced.2 This hyperuricaemia seen in association with pre-eclampsia is hypothesised to result from increased production,3,4 maternal renal dysfunction (reduced urate clearance or fractional excretion) and tissue ischaemia and acidosis.5 Outside pregnancy, hyperuricaemia has been linked through epidemiological studies to hypertension, metabolic syndrome, coronary artery disease, cerebrovascular disease, vascular dementia and chronic kidney disease.6

An association between maternal hyperuricaemia and poor pregnancy outcome was once considered a ‘truism’ in pre-eclampsia research7–10 but in our unit was felt to have become less reliable as neonatal care has advanced and pregnancy outcomes have improved.

One possible pathophysiological relationship with uric acid is blood volume. Circulating plasma volume is typically reduced in pre-eclampsia with a redistribution of extracellular fluid to the interstitial space.11 One driver for urate reabsorption by the kidney is volume contraction, either as a direct stimulus to the tubules or via increased angiotensin II production or sensitivity. This raises the question of whether a relationship exists between plasma volume contraction and serum uric acid in women with pre-eclampsia, and whether measuring both abnormalities would give greater ability to predict adverse maternal or fetal outcomes than measuring uric acid alone.

The purpose of this study was to determine if hyperuricaemia and haemoconcentration (measured indirectly as haematocrit or as haemoglobin concentrations) were associated with adverse pregnancy outcomes in a cohort of women referred for diagnosis and management of de novo hypertensive disorders of pregnancy.

Methods

We acquired data from two databases of hypertensive pregnant women, one of women referred to our Day Assessment Unit (DAU) and the other of all women referred for consultation to one of two renal physicians (MAB, GM) for management of hypertension in pregnancy (HIP). In general, all hypertensive pregnant women at our institution are referred to one or both of these services. Nonhypertensive women referred to the DAU for the management of type 1 diabetes were excluded unless they developed hypertension during their pregnancy, at which time they were included. Demographic features, clinical and laboratory data at the time of referral and delivery, and the outcome of mothers and babies are collected prospectively on each pregnancy primarily for the purpose of quality assurance. This database was established many years ago and we have used data from some of these women in other analyses.12

The databases were combined to eliminate duplication of women who were referred to both services. All hypertensive pregnancies, excluding those with chronic hypertension, white-coat hypertension (diagnosed by 24-hour ambulatory blood pressure monitoring) or renal disease, who were referred between 2000 and 2008, were included. Our complete, retrospective study population (n = 1880) included 355 pregnancies from the combined DAU and HIP databases, 1279 pregnancies only from the HIP database and 246 pregnancies only from the DAU database. In our Unit approximately 35% of our hypertensive pregnant women have gestational hypertension, 43% have pre-eclampsia, 16% have essential hypertension (17% of this cohort develop superimposed pre-eclampsia) and 6% have a primary renal or other secondary cause for hypertension. Sample size was calculated to assume a 20–40% prevalence of hyperuricaemia among our hypertensive pregnant population. We estimated a 10% predicted difference in clinically meaningful adverse outcomes between women with and without hyperuricaemia assuming a 20–40% frequency of adverse pregnancy outcome within our total hypertensive population as per similar studies.13 Based on an α of 0.01 and a β of 0.8, our calculated sample size per comparison group must be 50–300 to allow for two-sided analyses.

All women referred to the DAU are managed using a single protocol, which includes serum and urine laboratory investigations (haemoglobin, haematocrit, platelet count, electrolytes, creatinine, liver enzymes, uric acid, automated urinary dipstick and random urinary protein to creatinine ratio), four blood pressure measurements and cardiotocography. Similar laboratory investigations and blood pressure measurements are also completed for women referred for HIP consultation. Blood pressure was measured in the sitting position using a mercury sphygmomanometer. Ambulatory blood pressure or home blood pressure monitoring was not used. In general, women who are referred to the DAU tend to be further from delivery than women who are referred for HIP consultation.

The diagnoses of gestational hypertension and pre-eclampsia were made according to the diagnostic criteria of the Australasian Society for the Study of Hypertension in Pregnancy, as outlined in the International Society for the Study of Hypertension in Pregnancy (ISSHP) statement.14 Specifically, a blood pressure as an average in DAU or after overnight rest in hospital ≥140 mmHg systolic and/or 90 mmHg diastolic established the diagnosis of hypertension. Pre-eclampsia was diagnosed when hypertension was accompanied by proteinuria, defined as a random urinary protein to creatinine ratio >30 mg/mmol (95% of this cohort) or when hypertension and clinical evidence of maternal or fetal end-organ dysfunction was present in the absence of proteinuria (new onset maternal renal, hepatic or neurological disease, thrombocytopenia, or evidence of placental dysfunction including intrauterine growth restriction with or without reversed or absent uterine artery end-diastolic flow).12,15

The presence of hyperuricaemia was not a component in the diagnosis of pre-eclampsia nor in the decision to deliver any pregnancy. Gestational hypertension was diagnosed if there was hypertension only and no other signs or symptoms to suggest pre-eclampsia. Antihypertensive management, the decision to admit or continue follow up as an outpatient, and the indications for delivery and antihypertensive use were undertaken according to a common unit protocol that is closely adhered to in our institution and available at http://stgrenal.med.unsw.edu.au.

We tested the relationship between uric acid, haemoglobin, haematocrit and adverse pregnancy outcomes through the following analyses:

  •  Individual laboratory value for each of uric acid, haemoglobin and haematocrit as a continuous variable.
  •  The change (Δ) in laboratory value from the initial DAU value to the value closest to delivery using women referred to DAU only
  •  Δ uric acid ≥0.10 mmol/l (≥1.7 mg/dl) (n = 481)
  •  Δ haemoglobin ≥10 g/l (n = 537)
  •  Δ haematocrit ≥4% (n = 498).

To further investigate the relationship between serum uric acid and pregnancy outcome, we classified women according to the presence or absence of hyperuricaemia corrected for gestation, again using the uric acid value closest to delivery. These values were generally obtained within the 48 hours before delivery and no later than 1 week before. In this analysis, an elevated uric acid was defined as being one standard deviation above the gestation-specific mean16 as shown in Table 1. In addition, we used uric acid z-scores ([serum uric acid value − gestation specific mean15] / standard deviation of the population16) to account for gestation-specific alterations in uric acid and tested this as a continuous variable such that we could compare our results with those of previous studies.13

Table 1.   Values used to determine maternal serum hyperuricaemia based on uric acid (mmol/l) corrected for gestational age (mean +  1 standard deviation)16.
Gestational ageSerum uric acid (mmol/l)Serum uric acid (mg/dl)
<32 weeks>0.24>4.0
32–35 weeks>0.27>4.5
36–37 weeks>0.29>4.9
≥38 weeks>0.33>5.6

The primary outcome measures included composites of adverse maternal and adverse fetal outcomes that were predefined before statistical analysis. Adverse maternal outcome included one or more of: severe maternal hypertension after initial diagnosis (≥170/110 mmHg), renal insufficiency (creatinine >90 μmol/l; 1 mg/dl), liver disease (aspartate aminotransferase >40 U/l), cerebral irritation (defined as severe headaches with hyper-reflexia or sustained clonus ≥3 beats, repeated visual scotomata, or requiring magnesium sulphate) and thrombocytopenia (platelets < 100 × 109/l). Adverse fetal outcome included one or more of: perinatal death (any death after 20 weeks gestation until maternal discharge from hospital postpartum), small-for-gestational-age infant (SGA; <10th centile), admission to special-care nursery (criteria for admission included respiratory distress, low birthweight, hypoxia, jaundice and low blood sugar but did not include admission related to genetic disorders) and prematurity (<37 weeks of gestation).

Additional outcome measures recorded included delivery modality (caesarean or vaginal delivery) and requirement for two or more antihypertensive medications at discharge.

Data were analysed using stata version 11.0 (2009; Statacorp, College Station, TX, USA). The chi-square test and one-way analysis of variance were used to assess for differences in continuous data. Multivariable linear regression analysis was used to assess for the presence of a relationship using the continuous laboratory values and uric acid z-scores with outcome. All values met the assumptions of linear regression including linearity and normal error distribution. Logistic regression analysis was used to test the association between Δ laboratory value and outcome and also that between gestation-corrected hyperuricaemia and outcome. To test for a relationship between hyperuricaemia and haemoconcentration with final hypertensive diagnosis, adverse maternal outcome and adverse fetal outcome, correlation coefficients using pairwise deletion for missing data were calculated. Data are presented using odds ratios (OR) with 95% confidence intervals (95% CI), regression coefficients with t test scores or correlation coefficients. A value of < 0.01 was taken as an α level of statistical significance. This study was approved by the South Eastern Sydney and Illawarra Area Health Service Ethics Committee (HREC/09/STG/155).

Results

Linear regression analysis for uric acid closest to delivery, as a continuous variable, showed a significant association with final delivery diagnosis of pre-eclampsia (P = 0.001, 95% CI 0.01–0.04), composite adverse maternal outcome (P < 0.001, 95% CI 0.01–0.03), and composite adverse fetal outcome (P = 0.01, 95% CI 0.003–0.02). There was no association between hypertensive diagnosis or pregnancy outcome and either haemoglobin or haematocrit as continuous variables.

The percentage of women diagnosed with pre-eclampsia was significantly higher (56% versus 29%; < 0.001) if there was a 0.10 mmol/l (≥1.7 mg/dl) or more increase in plasma uric acid between that at the time of initial consultation and that closest to delivery (Δ uric acid). This trend was not observed for changes in haemoglobin (≥10 g/l) or haematocrit (≥4%). Logistic regression analysis showed a strong association between having pre-eclampsia as a final diagnosis and Δ uric acid (OR 4.6; P < 0.001) but no relationship was found for either composite adverse maternal or fetal outcomes with Δ uric acid. Similarly, Δ haemoglobin and Δ haematocrit were not associated with a final hypertensive diagnosis, composite adverse maternal outcome or composite adverse fetal outcome.

Linear regression analysis for gestation-specific uric acid z-score as a continuous variable showed a higher degree of association with final diagnosis of pre-eclampsia (P < 0.001, 95% CI 0.3–0.9), composite adverse maternal outcome (P < 0.001, 95% CI 0.4–0.7) and composite adverse fetal outcome (P < 0.001, 95% CI 0.6–1.0). Using the uric acid value closest to delivery, women with gestation-corrected hyperuricaemia, as defined by Table 1, were significantly more likely to be primiparous (68% versus 57%; P < 0.001), carry a twin pregnancy (10% versus 2%; P < 0.001), have higher systolic (143 versus 139 mmHg; P < 0.001) and diastolic (92 versus 90 mmHg; P < 0.001) blood pressure at time of physician consultation, be diagnosed with pre-eclampsia (62% versus 36%; P < 0.001), have or develop proteinuria (59% versus 38%; P < 0.001), deliver by caesarean section (53% versus 40%; P < 0.001) and require multiple antihypertensive medications postpartum (11% versus 4%; P < 0.001). Gestation-corrected hyperuricaemia was associated with less spontaneous labour (14% versus 23%; P < 0.001) and fewer cases of a previous diagnosis of hypertensive pregnancy (15% versus 24%; P < 0.001). There was no increased incidence of maternal diabetes in women with gestation-corrected hyperuricaemia.

Gestation-corrected hyperuricaemia was associated with a final diagnosis of pre-eclampsia (OR 2.0; 95% CI 1.3–3.0]; P < 0.01), composite adverse maternal outcome (OR 2.0; 95% CI 1.5–2.6; P < 0.001) and composite adverse fetal outcome (OR 2.2; 95% CI 1.7–3.0 P < 0.001). Multivariable analysis of pregnancy outcomes according to the absence or presence of gestation-corrected hyperuricaemia, with and without adjustments for parity, is presented in Table 2. Pregnancy outcomes associated with the presence of gestation-corrected hyperuricaemia include pre-eclampsia, maternal renal disease, SGA infant and < 37 weeks of gestation at delivery (but not severe prematurity defined as < 34 weeks of gestation). Additional linear regression analysis testing the relationship between serum uric acid and creatinine showed the two to be inter-related (P < 0.001). The mean creatinine was no different between gestational hypertensive women who remained so versus those who later went on to develop pre-eclampsia (63 μmol/l versus 64 μmol/l). Women who presented with pre-eclampsia had a higher mean level of creatinine than those who presented with gestational hypertension (69 μmol/l versus 63 μmol/l; P < 0.001) with both values being within the range considered normal for pregnancy.

Table 2.   Multivariable analysis of pregnancy outcomes*.
 Prevalence % (% missing)P adjusted**OR (95% CI) unadjustedOR (95% CI) adjusted**
  1. 95% CI, 95% confidence interval; GFR, glomerular filtration rate; NICU, neonatal intensive care unit; OR, odds ratio; SCN, special care nursery.

  2. *Multivariable analysis of pregnancy outcomes according to absence (OR 1.0) or presence of gestation-corrected hyperuricaemia (n = 1578; 16% missing urate values from study population of 1880), with and without adjustments for parity (parity adjusted model included all predictors as per unadjusted model) and prevalence of each outcome in women with gestation-corrected hyperuricaemia.

  3. **Adjusted for parity.

Pre-eclampsia62 (0)0.0002.3 (1.7–3.1)2.2 (1.7–3.0)
Severe maternal hypertension28 (2)0.0771.4 (1.0–2.0)1.4 (1.0–1.9)
Renal disease (reduced GFR)8 (26)0.0016.4 (2.2–18.4)6.4 (2.2–18.6)
Liver disease9 (25)0.1771.7 (0.9–3.2)1.6 (0.8–3.0)
Thrombocytopenia3 (1)0.2251.8 (0.7–4.7)1.8 (0.7–4.9)
Neurological disease7 (2)0.6660.9 (0.5–1.6)0.9 (0.5–1.6)
Magnesium sulphate use3 (0)0.2610.7 (0.3–1.6)0.6 (0.2–1.5)
SGA19 (11)0.0022.2 (1.4–3.3)2.0 (1.3–3.1)
Prematurity26 (0)0.0003.0 (2.0–4.6)2.6 (1.7–4.1)
<34 weeks gestation5 (0)0.1492.2 (0.7–6.8)2.3 (0.7–7.2)
NICU or SCN transfer12 (0)0.0970.7 (0.4–1.2)0.6 (0.4–1.1)

The prevalence of composite adverse maternal outcome, composite adverse fetal outcome and pre-eclampsia in women with normal serum uric acid (< 0.35 mmol/l; <5.88 mg/dl) was 40%, 41% and 42%, respectively (sensitivities 60%, 60%, 58%; specificities 60%, 56%, 63%). This tendency towards false positivity improved with the use of gestation-corrected uric acid, where normal values corresponded to a prevalence of composite adverse maternal outcome, composite adverse fetal outcome and pre-eclampsia of 26%, 22% and 28%, respectively (sensitivities 74%, 78%, 72%; specificities 51%, 49%, 63%). The positive and negative predictive values for gestation-corrected hyperuricaemia were 62% and 63% for composite adverse maternal outcome, 49% and 77% for composite adverse fetal outcome, and 72% and 55% for pre-eclampsia.

A key finding was that in women with gestational hypertension, who by definition are nonproteinuric and have no other maternal feature of pre-eclampsia, the presence of gestation-corrected hyperuricaemia was associated with an increased risk of composite adverse fetal outcome (OR 2.1; 95% CI 1.3–3.2; P < 0.01). Multivariable analysis within this population demonstrated SGA infant (OR 2.5; 95% CI 1.3–4.8; P < 0.01) and prematurity (OR 3.2; 95% CI 1.4–7.2) as the outcomes of significance (Table 3). Other findings in relation to uric acid as a continuous variable or as gestation-corrected hyperuricaemia are summarised in Table 3.

Table 3.   Key relationships between uric acid and maternal or fetal outcomes.
 Entire cohort (n = 1880)Gestational hypertension (n = 738)Pre-eclampsia (n = 1142)
Hyper-uricaemiaCorr hyper-uricaemiaHyper-uricaemiaCorr. hyper-uricaemiaHyper-uricaemiaCorr. hyper-uricaemia
  1. Corr. hyperuricaemia, gestation-corrected hyperuricaemia; GFR, glomerular filtration rate.

  2. Relationships are reported as odds ratio (OR) with 95% confidence interval (95% CI) in parentheses for the entire cohort and subdivided into women with gestational hypertension only (non-proteinuric and no other features of pre-eclampsia by definition) or pre-eclampsia.

  3. *Statistically significant at P < 0.01.

Final diagnosis of pre-eclampsia2.1 (1.6–2.7)*1.9 (1.3–3.0)*N/AN/AN/AN/A
Composite adverse maternal outcomes1.6 (1.3–2.0)*2.0 (1.5–2.6)*1.5 (1.0–2.3)1.4 (0.9–2.1)1.7 (1.2–2.2)*2.1 (1.5–2.9)*
Composite adverse fetal outcomes1.4 (1.1–1.8)*2.2 (1.7–3.0)*1.4 (0.9–2.2)2.1 (1.3–3.2)*1.4 (1.0–1.9)2.4 (1.7–3.4)*
Renal dysfunction (reduced GFR)5.6 (2.4–13.0)*6.0 (2.1–17.3)*2.8 (0.7–11.3)1.8 (0.4–7.4)4.5 (2.4–8.5)*11.5 (3.6–37.2)*
Maternal thrombocytopenia5.5 (1.6–19.6)*3.4 (0.9–12.5)2.8 (0.5–17.4)1.9 (0.3–12.3)2.6 (1.0–6.5)2.1 (0.7–6.2)
SGA (<10th centile)1.4 (1.0–2.0)2.0 (1.3–3.0)*1.4 (0.8–2.7)2.5 (1.3–4.8)*1.4 (1.0–1.9)1.8 (1.2–2.7)*
Prematurity (<37 weeks)1.4 (1.0–2.0)3.1 (2.0–4.7)*1.5 (0.7–3.2)3.2 (1.4–7.2)*1.4 (1.1–1.9)3.8 (2.6–5.6)*

In an attempt to determine if uric acid was correlated in any way to either haemoglobin or haematocrit, we calculated correlation coefficients by final hypertensive diagnosis and pregnancy outcome. No correlation was found between uric acid or uric acid z-score and haemoglobin for final hypertensive diagnosis or pregnancy outcome. Correlation was slightly better, although still poor and nonsignificant, between uric acid z-score and haematocrit. Figure 1 shows the nonsignificant correlation coefficients for uric acid z-score and haematocrit for each of final hypertensive diagnosis (gestational hypertension or pre-eclampsia), composite adverse maternal outcome and composite adverse fetal outcome.

Figure 1.

 Scatterplots showing the lack of association between uric acid z-score (UA Z-s) and haematocrit by outcome; final diagnosis: gestational hypertension (a), final diagnosis: pre-eclampsia (b), composite adverse maternal outcome (c) and composite adverse fetal outcome (d). The Pearson product-moment correlation coefficient (CC) and sample size (n) for each association are shown.

Discussion

This study shows that maternal hyperuricaemia, particularly gestation-corrected hyperuricaemia, measured near delivery is associated with adverse maternal and fetal outcomes. Gestation-corrected hyperuricaemia is associated with an increased prevalence of SGA infants and prematurity, perhaps surprisingly even in women who have only gestational hypertension without proteinuria or any other maternal feature of pre-eclampsia. This observation suggests that gestational hypertension in the presence of hyperuricaemia is a disease with increased fetal risk.

Previous studies have shown a relationship between hyperuricaemia and adverse obstetric outcome in hypertensive pregnancy.7–10,13,17,18 Our data provide further support of this relationship and we show that in women with hypertensive pregnancy, the presence of elevated uric acid (particularly gestation-corrected hyperuricaemia) is associated with a diagnosis of pre-eclampsia more so than with gestational hypertension. Hyperuricaemia in women with pre-eclampsia identifies a population prone to develop adverse maternal and fetal outcomes, specifically maternal renal disease, SGA and preterm birth. Furthermore, hyperuricaemia in women with gestational hypertension alone is associated with adverse fetal outcomes including SGA and preterm birth, suggesting that this subpopulation of gestational hypertensive women can no longer be considered to have a ‘benign’ hypertensive disorder of pregnancy.

In our study, the increased incidence of preterm birth coincided with a reduction in the percentage of women achieving spontaneous labour, which almost certainly represented the need for expedited delivery because of the severity of maternal or fetal hypertensive disease. It should be noted that neither plasma uric acid, nor a change in uric acid, has ever been an indication for delivery in our unit. Alongside the increased incidence of preterm birth was an increased SGA rate. These effects were shown to be further related to uric acid concentration as the risk of adverse fetal outcome increased, with increasing range of uric acid beyond 0.35 mmol/l (5.9 mg/dl) and with gestation-corrected hyperuricaemia. This was true not just for the combined cohort of women with pre-eclampsia or gestational hypertension but also when restricted to women with gestational hypertension alone. It is improbable that severity of maternal hypertension led to premature delivery in the gestational hypertension group as the incidence of severe maternal hypertension was only 16%, compared with 33% among women with pre-eclampsia.

Similarly, the risk of adverse maternal outcome increased with increasing uric acid. Specifically, an elevated uric acid level corrected for gestation was associated with a five-fold increased risk of development of maternal renal dysfunction as measured by increased serum creatinine when compared with the risk in the absence of gestation-corrected hyperuricaemia. Our analyses suggest that serum uric acid and serum creatinine were strongly inter-related; however, the majority of women in our population had a serum creatinine within what is considered a normal range for pregnancy, suggesting that serum creatinine is not a sensitive marker of progressive hypertensive disease in pregnancy.

The aetiology of hyperuricaemia in pre-eclampsia is thought to be multifactorial, with contributions from decreased renal tubular secretion, increased oxidative stress as the result of placental ischaemia and maternal reduction in glomerular filtration rate.5 Our results suggest that maternal renal function is an important factor leading to hyperuricaemia but our study design did not allow us to further investigate this relationship.

In a nested case–control study of 972 women, hyperuricaemia in the setting of gestational hypertension was associated with shorter gestation and reduced birthweight even in the absence of proteinuria.13 The authors concluded that uric acid should be considered at least as important as proteinuria in pregnancies complicated by gestational hypertension. We have previously shown that proteinuric pre-eclamptic women with hyperuricaemia at the time of delivery were more likely to have preterm birth than other women with clinical features of pre-eclampsia who did not have proteinuria.12 In a recent systematic review, elevated serum uric acid was associated with an almost doubled risk of severe maternal complications in pre-eclamptic women; however, the authors concluded that uric acid measurement is not a clinically useful test because of a below expected positive likelihood ratio.17 A subsequent bi-variate meta-analysis looking at multiple uric acid cutoff values suggested that uric acid measurement is clinically useful in predicting maternal complications of pre-eclampsia.18 Both of the above studies were limited by the small sample sizes available.

Plasma volume contraction is a common finding in pre-eclamptic pregnancies.19 An elevation in haemoglobin or haematocrit is hypothesised to be the result of a combination of reduced plasma circulating volume and enhanced erythropoiesis because of underlying placental hypoxia.20 Erythrocyte parameters in normal and pre-eclamptic pregnancies have been shown to differ,21 and both haemoglobin and haematocrit have been included in multivariable prediction schemes, although neither seems to have been compared against maternal or fetal outcome.22 Our results did not demonstrate any association between our measures of haemoconcentration and hypertensive diagnosis, obstetric outcome or uric acid. This implies that additional factors beyond volume contraction carry a larger responsibility for uric acid elevation in hypertensive pregnancy or that haemoglobin and haematocrit are very poor markers of volume contraction in pregnancy. Whichever explanation is correct, this decreases the importance of repeatedly measuring haemoglobin and haematocrit in hypertensive pregnancy and implies that both tests are unreliable in identifying women at higher risk of adverse maternal and fetal outcomes.

High values of uric acid in the setting of renal disease are not surprising as uric acid clearance is largely reliant on renal excretion.23 Renal dysfunction is a common finding in women with pre-eclampsia and is characterised by glomerular endothelial injury that results in a reduction in ultrafiltration capacity.24 Moreover, Angiotensin II stimulates renal urate reabsorption resulting in less urate excretion25 and evidence suggests that pre-eclampsia is a disorder of enhanced sensitivity to angiotensin II.26

We believe that it is important to adjust measured uric acid levels for gestational age because uric acid concentration is influenced by the ongoing physiological changes that occur throughout pregnancy. In early pregnancy, plasma uric acid falls by 25–35%, primarily because of increased blood volume and glomerular filtration rate.27 The concentration begins to rise mid-pregnancy and approximates or even surpasses values in non-pregnant women by term.16,28 Using a uric acid z-score to account for the gestation-related changes that occur in uric acid concentration, our results showed that both maternal and fetal outcome, and final hypertensive diagnosis are related to uric acid concentration at the time of delivery. Unfortunately, the use of z-scores in clinical practice is impractical because they require calculation, which would increase the potential for error. We therefore tested the utility of a gestation-corrected uric acid cutoff value that was based on one standard deviation above the mean16 (Table 1) because this can be incorporated easily into clinical practice. We found similarly that this practical measure related strongly to adverse maternal and fetal outcomes.

Plasma uric acid concentration tends to fall initially then rise as pregnancy progresses.27 It has been suggested that abrupt increases in uric acid concentration may be related to the development of an underlying hypertensive disease.18 This is consistent with our observation that women who had a Δ uric acid of ≥10 mmol/l (≥1.7 mg/dl) were more likely to develop pre-eclampsia than those who did not. Despite this, our study did not find an association between the absolute rise in plasma uric acid and adverse maternal or fetal outcomes.

There are limitations to this study. Our database population is comprised entirely of women referred for specialist physician care during their pregnancy, perhaps excluding women with less severe disease. We were limited to the data contained within the databases, which made it impossible to determine the gestation at which a hypertensive diagnosis was first made. As such, we were unable to subclassify women with pre-eclampsia as early (≤34 weeks of gestation) or late (>34 weeks of gestation). In addition, the speed at which pre-eclampsia developed was not accounted for in our databases and so we could not compare gradual versus abrupt onset pre-eclampsia. For most women only laboratory values closest to delivery were available and as such, we were unable to determine the temporal relationship between uric acid and hypertensive diagnosis or initiation of antihypertensive medications. In addition, uric acid measurement did not occur at a uniform time during each pregnancy and so the gestation at which this was measured varied by woman. Second, intravascular fluid administration in hospital was not recorded in the database, which if given before phlebotomy, would have affected the measures of haemoconcentration. However, giving intravenous fluids for the antenatal management of pre-eclampsia is not usual practice in this unit, and is unlikely to be a major factor influencing our findings. Other factors that could impact our measure of haemoglobin and haematocrit including maternal iron status, haemoglobinopathies, and disorders of haemolysis both related and unrelated to pre-eclampsia were not recorded in the database. Although uric acid concentration is not routinely considered a criterion for diagnosing pre-eclampsia or used in management decisions regarding hypertensive women in our Unit, care providers would have been aware of these levels and therefore may have impacted timing of delivery for some women. It could not however have impacted the finding that gestation-corrected hyperuricaemia was associated with higher SGA rates, even in women with gestational hypertension alone. This lends some support to the observation that uric acid may affect mechanisms relevant to fetal growth.29,30 Lastly, the prevalence of many of our individual maternal outcomes was low and therefore the power to detect differences in these secondary analyses was limited by small sample size. This was exacerbated by a relatively high percentage of missing values for these variables (maternal renal and liver disease).

Conclusion

Our data suggest that the presence of hyperuricaemia, but not elevated haemoglobin or haematocrit, identifies a population of hypertensive pregnant women at increased risk of adverse maternal and fetal outcome. Hyperuricaemia in the setting of pre-eclampsia was associated with both maternal and fetal adverse outcomes. Hyperuricaemia in the setting of gestational hypertension (non-proteinuric by definition) was associated with adverse fetal outcome, specifically SGA and prematurity, suggesting that this subpopulation of gestational hypertensive women cannot be considered to have a benign hypertensive disease. Hyperuricaemia can be based on a cutoff level above 0.35 mmol/l (>5.9 mg/dl), but use of a gestation-adjusted uric acid level provides a stronger association with obstetric outcome. Calculating the serum uric acid rise (Δ uric acid) may be useful in predicting which women will go on to develop pre-eclampsia.

On the basis of these findings we believe that hyperuricaemia is a component of the pre-eclamptic syndrome and that the measurement of uric acid concentration remains helpful in the management of high-risk hypertensive pregnancies (without primary renal disease) as it is associated with an increased likelihood of developing maternal and fetal complications. We do not have data to show that this is any more or less important than blood pressure levels, maternal symptoms, or proteinuria in predicting adverse pregnancy outcomes but believe it should remain an integral clinical measure in the assessment of hypertensive pregnant women. Future research is needed to determine if uric acid itself is pathogenic, whether it should be included in the diagnostic criteria for pre-eclampsia and whether gestation-corrected plasma uric acid should be used to direct management.

Disclosure of interests

We have no conflicts of interest to declare.

Contribution to authorship

TLH and MAB wrote the study protocol, performed the analyses and wrote the first draft of the manuscript. JMR, GJM and GKD made substantial contributions to study design and manuscript revision.

Details of ethics approval

This study was reviewed and approved by the South Eastern Sydney and Illawarra Area Health Service Ethics Committee (HREC/09/STG/155).

Funding

This study was funded without external grant funding.

Acknowledgements

We have none to declare.

Ancillary