Acute Renal Failure During Eclampsia: Incidence Risks Factors and Outcome in Intensive Care Unit

Abstract

Objective: To assess the incidence, risk factors and the outcome of acute renal failure (ARF) associated with eclampsia in intensive care unit (ICU). Design: Prospective and analytic study. Setting: A surgical ICU in a university hospital. Patients: 178 consecutive women with eclampsia admitted to an intensive care unit during seven years. ARF was defined by a serum creatinine concentration > 140 µmol/L. Results: The incidence of ARF was 25.8%. In univariate analysis the severity of patient illness, the complications associated with eclampsia (disseminated intravascular coagulation, Hellp syndrome, neurologic complications, abruptio placenta, aspiration pneumonia, delivery hemorrhage) were significantly associated with ARF. In a logistic regression model, risk factors for ARF included organ system failure (OSF) odds ratio (OR) = 1.81 confidence interval (CI) [1.08–3.05], bilirubin > 12 µmol/L OR = 4.42 CI [1.54–12.68], uric acid > 5.9 g/dL OR = 16.5 CI [3.09–87.94], abruptio placenta OR = 0.2 7 CI [0.08–0.99], and oliguria OR = 0.10 CI [0.03–0.44]. In contrast, severity of blood pressure or proteinuria on dipstick were not associated with ARF. However, in this series, 15 women required dialysis in the short term and one required long‐term dialysis. ARF associated with eclampsia was significantly associated with mortality (32.6% versus 9.1% p = 0.0001). Conclusion: ARF with eclampsia is a frequent situation that required intensive management when risks factors were present. The need for dialysis was a rare condition.

• Acute renal failure
• Eclampsia
• Hellp syndrome
• Disseminated intravascular coagulation

Introduction

Eclampsia remains an important cause of maternal and fetal morbidity and mortality in developing countries.[1], [2] Among the complications associated with preeclampsia–eclampsia, incidence of acute renal failure (ARF) varies from 4% to 23%.[3], [4] This variability is due to different factors such as the place of study and the severity of preeclampsia. Preeclampsia–eclampsia constitutes the first cause of the ARF in pregnancy in many countries.[5], [6], [7] Some risk factors for developing ARF during preeclampsia–eclampsia were perfectly identified, such as abruption placenta or Hellp syndrome.[4] Other causes would be implied and would be found in glomerular endotheliosis, a favorable bed for the occurrence of the ARF. Despite the considerable progress of intensive care and the early management of these patients, the evolution may require temporary or definitive hemodialysis.[4] The aim of this study is to determine the incidence, the risk factors and the outcome of ARF associated with eclampsia.

Materials and Methods

Between January 1, 1993 and December 31, 2000, 178 patients with the diagnosis of eclampsia were managed after delivery at the intensive care unit of University Medical Center. Data were collected prospectively. Eclampsia was defined as the occurrence of seizures in the presence of preeclampsia (shown by hypertension diastolic blood pressure of at least 90 mmHg, proteinuria one “plus” or at least 0.3 g/24 h occurring after 20 weeks gestation). Patients with any cause for convulsion other than eclampsia were excluded. Acute renal failure was defined by the following criteria: a serum creatinine concentration > 140 µmol/L at the time of admission in intensive care unit without preexisting renal disease.

Physiologic variables were measured at the time of ICU admission to calculate severity scoring indexes: Acute Physiology and Chronic Health Evaluation System II (APACHE II) score, the Simplified Acute Physiology Score (SAPS), the number of organ system failure (OSF), and Simplified Acute Physiologic Score applied to obstetric (SAPSO).[8], [9], [10], [11] The following variables were recorded at admission in hospital: demographic data: age, weight, mean gestational age, systolic and diastolic blood pressure before any therapeutic, gravidity, clinical findings; abdominal pain, nausea and vomiting, jaundice, edema, proteinuria by dipstick, laboratory findings (platelet count, hemoglobin, bilirubin uric acid concentration, proteinuria, aspartate aminotransferase and lactate dehydrogenase values, prothrombin time, fibrinogen, blood urea nitrogen and creatinine serum concentration. Urine was also analyzed for proteinuria on admission and a 24‐hour urine collection was performed to determine total protein clearance. Maternal complications were noted: unconscious patients, recurrent seizures or focal neurologic signs were investigated by cerebral computed tomographic scan to search neurologic complication as intracerebral hemorrhage, ischemia or cerebral edema. Diagnosis of Hellp syndrome required the following laboratory findings: hemolysis defined by abnormal peripheral smear, increased bilirubin (> 12 µmol/L) and increased lactic dehydrogenase (LDH > 600 U/L), elevated liver enzymes defined as increased serum aspartate aminotransferase (ASAT) level > 70 IU/L, alanine aminotransferase (ALAT) level > 60 IU/L and low platelets defined as platelet less than 100,000 mm^{− 3}.[12] Diagnosis of disseminated intravascular coagulation (DIC) included: low platelets (100,000 mm^{− 3}), fibrinogen < 300 mg/dL and prolonged prothrombin time > 14 seconds. Abruptio placentae were diagnosed by inspection of the placenta at delivery. Diagnosis of pulmonary edema included acute onset of shortness of breath, hypoxia, and radiologic evidence of pulmonary edema. Oliguria was defined by less than 400 ml urine/24 h.

All women were managed with a standardized protocol. The patient received intravenously, diazepam 10 mg boluses or midazolam 5 mg repeated as required to stop convulsion. The airway was secured by intratracheal intubation. The most common antihypertensive drugs used are nicardipine, hydralazine or clonidine. Fluid balance was managed by monitoring central venous pressure and urine output. Indications for dialysis were volume overload, hyperkaliemia (above 7 mEq/L), serum creatinine concentration > 700 µmol/L and severe metabolic acidosis (bicarbonate under 15 mEq/L).

Two subsets of patients were identified according to the presence (ARF +) or absence (ARF −) of acute renal failure. Comparisons between groups were made with the unpaired t test for continuous variables and the chi square test for categorical data with calculation of the relative risk and its 95% confidence interval. p < 0.05 was considered significant. Stepwise logistic regression was used to explore the effect of several variables on risk factors for ARF.

Results

A total of 178 women had eclampsia. 46 patients with ARF met the inclusion criteria (25.8% of total ICU admission). Patients with preexisting diseases were excluded. Mean serum creatinine concentration was 343 ± 220 µmol/L on admission in ICU and blood urea nitrogen serum concentration was 1.3 ± 0.7 mg/dL.

Both groups were similar with regard to age, weight, gravidity and gestation at delivery (Table 1). The duration of hospitalization was longer in the group with ARF (8.7 days vs. 4.7 days p < 0.006). There was no difference in the occurrence of ARF when the convulsions occurred before delivery or after delivery. ARF was not associated with number of convulsions (4.2 ± 1.5 versus 3.5 ± 1.4 in patients without ARF p = 0.064).

Table 1. Demographic Data

	ARF + (n = 46)	ARF − (n = 132)	p
Mean age (y)	28.4 ± 7.3	26.4 ± 7.2	NS
Weight (Kg)	67 ± 12	68 ± 13	NS
Gestation at delivery (wk)	34.4 ± 3.9	35.6 ± 3.8	NS
Primigravid (%)	24 (52%)	88 (68%)	NS
Hospital stay (day)	8.7 ± 5.4	4.7 ± 3.4	0.006
Number of convulsions	4.2 ± 1.5	3.5 ± 1.4	NS
Time of convulsions
Antepartum	34 (73.9%)	101 (76.5%)	NS
Postpartum	12 (26.1%)	31 (23.5%)

Comparison of eclampsia with and without ARF. Values are mean ± SD or percent.

The severity of patient illness at time of admission is shown in Table 2. At the time of ICU admission, the patients in group ARF + were more severely ill than those in the other group. The mean SAPS I, SAPSO, APACHE II and the number of OSF (including renal dysfunction) were significantly higher when eclampsia was associated with ARF. However, Glasgow Coma Scale was not a very good predictive variable for ARF. The following complications were significantly associated with ARF: DIC, Hellp syndrome, abruptio placenta, aspiration pneumonia and delivery hemorrhage. However, pulmonary edema was not associated with ARF in this study. ARF was associated significantly with mortality (Table 3).

Table 2. Severity Scoring at Admission in Eclamptic With and Without ARF

	ARF + (n = 46)	ARF − (n = 132)	p
Glasgow Coma Scale	10.4 ± 3.1	10.7 ± 3	NS
SAPS	12.8 ± 4.7	8.4 ± 3.4	0.0001
SAPSO	17.9 ± 6.2	10.6 ± 3.9	0.0001
OSF	1.3 ± 0.3	1 ± 0.7	0.00001
APACHE II	16.9 ± 4.8	11.7 ± 3.9	0.0001

Values are mean ± SD.

Table 3. Complications Asssocied with ARF in Eclamptic Women

	ARF + (n = 46)	ARF − (n = 132)	RR	95% CI	p
DIC	17 (37%)	14 (10.6%)	4.94	2.18 − 11.17	0.0001
Hellp Syndrome	17 (37%)	15 (11.4%)	4.57	2.04 − 10.22	0.0001
Neurologic complications	13 (28.3%)	13 (9.8%)	3.6	1.52 − 8.52	0.002
Abruptio placenta	13 (28.3%)	12 (9.1%)	3.93	1.64 − 9.64	0.001
Pulmonary edema	3 (6.5%)	10 (7.6%)	0.85	0.22 − 3.23	NS
Aspiration pneumonia	8 (17.4%)	9 (6.8%)	2.87	1.03 − 7.97	0.036
Delivery hemorrhage	11 (23.9%)	11 (8.3%)	3.45	1.38 − 8.64	0.006
Maternal death	15 (32.6%)	12 (9.1%)	4.83	2.05 − 11.38	0.000

Values are percent.

Clinical findings at admission are reported in Table 4. Jaundice and nausea and vomiting were more frequent in group with ARF in comparison with group without ARF (45.7% vs. 19.7% and 43.5% vs. 13.6% p < 0.001 and p < 0.0001 respectively). However, there is no difference between the two groups concerning hemodynamics parameters at admission or protein in urine dipstick. There was no association between ARF and the presence or absence of edema. Oliguria was a good predictor of ARF when it was present at admission in ICU [relative risk (RR) 33,1; confidence interval (CI) 9.15–119.5; p < 0.0001]. In univariate analysis, bilirubin serum concentration > 12 µmol/L, prothrombin time < 50%, fibrinogen serum concentration < 0.3 g/dL, hemoglobin < 9 g/dL, platelet count < 100,000/mm3 and uric acid > 5.9 mg/dL were associated with ARF. There was no association between serum albumin concentration, aspartate aminotransferase, lactate dehydrogenase and quantitative proteinuria (Table 4). Assisted respiration was observed similarly in the two groups.

Table 4. Clinical Features and Laboratory Data for Patients With and Without ARF

	ARF + (n = 46)	ARF − (n = 132)	RR	95% CI	p
Clinical features
Nausea and vomiting	21 (45.7%)	26 (19.7%)	3.42	1.66 − 7.04	0.001
Jaundice	20 (43.5%)	18 (13.6%)	4.87	2.26 − 10.48	0.0001
Abdominal pain	19 (41.3%)	25 (18.9%)	3	1.45 − 6.25	0.002
Oliguria	20 (43.5%)	3 (2.3%)	33.1	9.15 − 119.5	0.0001
Proteinuria on dipstick
0 +, 1 +, 2 +	43 (93.5%)	111 (84.1%)	2.71	0.76 − 9.55	NS
3 +, 4 +	23 (50%)	69 (52.3%)	0.91	0.46 − 1.78	NS
Diastolic blood pressure
≤ 90 mmHg	3 (6.5%)	11 (8.3%)	0.76	0.20 − 2.88	NS
91–109 mmHg	16 (34.8%)	57 (43.2%)	0.70	0.35 − 1.41	NS
≥ 110 mmHg	27 (58.7%)	64 (48.5%)	1.51	0.76 − 2.97	NS
Edema
Absent	11 (23.9%)	19 (14.4%)	1.86	0.81 − 4.30	NS
Generalized	28 (60.9%)	87 (65.9%)	0.80	0.40 − 1.60	NS
Localised	7 (15.2%)	26 (19.7%)	0.73	0.29 − 1.82	NS
Laboratory data
Bilirubin > 20.5 µmol/l	22 (47.8%)	20 (15.2%)	5.13	2.42 − 10.85	0.000
Prothombin time < 50%	14 (30.4%)	6 (4.5%)	9.19	3.27 − 25.8	0.000
Fibrinogen < 3 g/l	22 (47.8%)	38 (28.8%)	2.26	1.14 − 4.52	0.019
Platelet count < 100,000/mm3	29 (63%)	32 (24.2%)	5.33	2.59 − 10.94	0.000
Uric acid > 5.9 mg/dl	44 (95.7%)	72 (54.5%)	18.33	4.26 − 78.77	0.000
Hemoglobin < 9 g/dl	19 (41.3%)	77 (58.3%)	0.5	0.25 − 0.99	0.046

At admission in ICU. Values are mean ± SD or percent.

Diuretic use, mainly furosemide and low dose dopamine < 3 mcg/kg/min use were more frequent in patient with ARF p < 0.0001. Mean duration of diuretic use and dopamine use were longer in patients with ARF compared to patients without ARF p < 0.001. Mode of delivery was not associated with ARF (Table 5).

Table 5. Therapeutique Use and Mode of Delivery

	ARF + (n = 46)	ARF − (n = 132)	RR	95% CI	p
Furosemide use (%)	40 (87%)	66 (50%)	6.67	2.65 − 16	0.0001
Duration (days)	4.92 ± 3.8	2.4 ± 1.9			0.001
Dopamine use (%)	25 (54.3%)	7 (5.3%)	21.26	8.16 − 55.36	0.001
Duration (days)	5.7 ± 4.2	2.7 ± 1.4			0.009
Assisted respiration (%)	41 (89.1%)	112 (84.8%)	1.46	0.51 − 4.15	NS
Duration (days)	5.8 ± 5.5	2.3 ± 2.3			0.014
Cesarean delivery (%)	29 (63%)	88 (66.7%)			NS

Values are mean ± SD or percent.

15 patients during this period required hemodialysis. The average number of dialysis procedures per patient was 3 (range 1 to 8). The recovery of diuresis was observed on 1.4 ± 2.7 days (extremes 0 to 12 days) and complete recovery of renal function was observed on 14 ± 8 days. Only one patient required permanent dialysis for chronic renal failure. She had a kidney biopsy that revealed cortical necrosis.

Using a stepwise multiple logistic regression model, organ system failure, bilirubin serum concentration > 1.2 mg/dL, uric acid concentration > 5.9 mg/dL, abruptio placenta and oliguria were independently associated with acute renal failure (Table 6).

Table 6. Multivariate Predictors of Acute Renal Failure Associed with Eclampsia

Variable	Beta	Odds ratio	95% CI	p
Constant	− 1.186	–	–	–
OSF	0.596	1.81	1.08 − 3.05	0.024
Bilirubin > 20.5 µmol/l	1.486	4.42	1.54 − 12.68	0.006
Uric acid > 5.9 mg/dl	2.804	16.5	3.09 − 87.94	0.001
Abruptio placenta	− 1.288	0.27	0.08 − 0.99	0.049
Oliguria	− 2.250	0.10	0.03 − 0.44	0.002

Discussion

The incidence of ARF in this study is still high (25,8%) compared to that found in developed countries (4%) but identical to that found by Selcuk in Turkey (19.3%).[7], [13], [14] It constitutes the main cause of obstetric ARF in our context.[15] The difference in the incidence is explained by the definition of ARF. Some authors included only the patients requiring hemodialysis, others, patients with oliguria, others, patients with creatinine level > 105 µmol/L. Others remain evasive on the definition of ARF. The threshold of 140 µmol/L was selected in an arbitrary way and corresponds to what is usually retained in the literature.[16], [17], [18] The second reason is that the majority of the studies focused on the ARF in the pre‐eclamptic patient, thus, excluding the severe forms like eclamptic patients with numerous lesions. Moreover, the place of the study (intensive care or gynecoobstetric service) as well as the country plays a considerable role in the incidence of this complication. In developing countries, the incidence of eclampsia is 10 times higher than that observed in the developed countries and therefore, complications are higher. Furthermore, as it is the case of our hospital, it constitutes a center of reference for the assumption of responsibility of these patients with intensive care and the possibility of performing hemodialysis.[1] Many patients are transferred secondarily in our hospital and their transportation is not medically equipped. Many authors insist on the fact that the term and the time of the occurrence of convulsions lead to complications.[13] A term lower than 32 SA is one of the factors that cause ARF. This is not the case when it comes to the time when convulsions happen, the level of blood pressure and the importance of the proteinuria or the edemas. On the other hand, the existence of a jaundice of nauseous vomiting or abdominal pains via the hypovolemia worsens the renal hypoperfusion. This symptomatology is often synonymous with Hellp syndrome.[19] In a recent study, Drakeley analyzing a number of ARF in pre‐eclamptic patients has shown that 50% of the parturient presented a Hellp syndrome.[16] A term < to 32 is associated with the occurrence of abruptio placenta for Mattar.[13] The abruptio placenta is one of the main factors that lead to the occurrence of ARF. The hemorrhage accompanying the abruptio placenta as well as the DIC frequently associated, support the release of mediators, the vasoconstriction related to separation, the state of shock with renal hypoperfusion, ischaemia and acute tubular necroses. The abruptio placenta, in the absence of pre‐eclampsia/eclampsia, is not associated with important renal lesions.[4]

By the same mechanism, the hemorrhage during delivery following the example of all hypovolemias would make the occurrence of the ARF more likely. The glomerular endothelioses with ballonisation and vacuolization of the endothelial glomerular cells would constitute a lesion of renal hypoperfusion.[20] The various pilot scores, sign of the seriousness of the illness of the patients as they are admitted to ICU are significantly associated with the occurrence of ARF. The majority of these scores take into account the values of the creatinine. Score OSF seems the most predictable because the deterioration of the renal function would be related to a failure of the renal blood flow.

The evolution of this ARF is, as a general rule, favorable if the treatment is aggressive as soon as the patients are treated. This goes through an optimal filling using the physiological salt solution or the blood derivatives in the event of disorder of the blood crase. This filling must be made under a strict monitoring of central venous pressure even if this one constitutes a bad reflection of the preload. The catheterism of Swan Ganz must be used in oligoanuric forms or in case of pulmonary edema. The recourse to diuretics must be large after the deliveries just as the recourse to the dopamine whose selective vasodilator effects improve the renal perfusion.[21], [22] The bad management of ARF at its early stages leads to cortical necrosis. It must be evoked after 3 to 4 weeks of hemodialysis and must lead to the renal biopsy.[4] In general, the return to a normal renal function is the outcome after 3 or 4 hemodialysis in the pure cases of preeclampsias.

In conclusion, ARF is a complication frequently associated with eclampsia. The DIC, the abruptio placenta, the Hellp syndrome as well as all the situations leading to a hypovolemia can break the fragile balance of the pre‐eclamptic patient's kidney. An energetic and early treatment prevents, in the majority of cases, the recourse to the hemodialysis and even to the evolution towards cortical necrosis.