Obstetric critical care: A blueprint for improved outcomes

Gerda G. Zeeman, MD, PhD

Pregnancy induces a multitude ofrather profound physiologic he-modynamic alterations. Amongthese are substantive increases

in total blood volume, cardiac output, anduterine blood flow (1). The major cardio-vascular and hematologic changes ofpregnancy, and the particular pathologiesunique to pregnancy, most importantly,(pre)eclampsia, contribute additional chal-lenges and require additional understand-ing. Certain medical conditions present dif-ferently during pregnancy, and varioustherapies may have their effect on the de-veloping fetus and may influence the planfor delivery. In addition, there are severalacute circumstances, such as massive ob-stetric hemorrhage, specific to pregnancythat require a high degree of expertise foreffective management.

Increasing numbers of women with un-derlying chronic medical conditions are

now able to conceive and carry a fetus toterm. For instance, women with longstand-ing diabetes mellitus or complex congenitalcardiac defects are now counselled muchmore liberally when they present with thedesire to raise a family. Such women obvi-ously require a greater degree of medicalcare throughout the antepartum and post-partum periods. Management of suchwomen throughout labor and delivery mayrequire a period of observation in an inten-sive care unit (ICU). Knowledge of the spe-cial circumstances in which these womenmay present is essential for both the obste-tricians and the intensivists who are re-sponsible for their care. Through review ofthe current literature and an update on thepathophysiology of (pre)eclampsia andtreatment of massive obstetric hemor-rhage, this review aims to update the inten-sivist who is confronted with the obstetricpatient.

Overview

Admission of an obstetric patient to anICU in developed countries occurs inroughly 2–4/1,000 deliveries (Table 1).Despite this low prevalence, the overallacuity of this group of patients is high.Perinatal ICU services vary widely from

the dedicated obstetric ICU (2, 3) to thetransfer of a critically ill obstetric patientto a medical/surgical ICU (4–8) or to anICU in another hospital (9).

During the last 15 yrs, several reportsfrom a variety of centers all over theworld have described the characteristicsand treatment of critically ill pregnant orpuerperal women (Table 1). The studiesreport significant variations in patientpopulations, definition of major morbid-ity, ICU admission criteria, usage rates,outcomes, and treatment. In addition,differences in hospital settings, nursingpolicies, and management protocols mayadd to the observed variations. Differ-ences in access to health care in the citedstudies make comparisons of standard ofcare and recommendations for improve-ment difficult.

Usually, women are young and fre-quently primipara. Most admissions, there-fore, occurred in low-risk pregnancieswithout contributory obstetric history. Ingeneral, for most obstetric patients, rapidrecovery followed correction of the acuteinsult. This is reflected in that most womenwere admitted to an ICU for �48 hrs. Ex-pectedly, this is shorter than the meanlength of stay in the nonpregnant popula-tion. The most frequent admission reason

From the Department of Obstetrics and Gynecol-ogy, University Medical Center, Groningen, The Neth-erlands.

The author has not disclosed any potential con-flicts of interest.

Copyright © 2006 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins

DOI: 10.1097/01.CCM.0000231884.99763.69

Introduction: Obstetric patients are generally young andhealthy. However, the potential for catastrophic complications isreal, and despite the therapeutic advances of the last few de-cades, maternal morbidity and mortality continue to occur. Thismay be related to the pregnancy itself, aggravation of a preex-isting illness, or complications of the (operative) delivery.

Purpose: The purpose of this review is two-fold: first, to providean update on currently available reports pertaining to importantcritical care issues of the obstetric patient population and, second, topresent current comprehensive treatment options for preeclampsiaand massive obstetric hemorrhage because both are responsible forthe majority of maternal mortality and morbidity worldwide.

Results: The most common reasons for intensive care unitadmission are hypertensive disorders and massive obstetric hem-orrhage. Timely delivery and prompt initiation of antihypertensivetherapy for severe hypertension form the mainstay of care inpreeclampsia. Restoration of circulating blood volume and rapid

control of bleeding and impaired coagulation are the main factorsin the management of massive obstetric hemorrhage. Puerperalmorbidity has become the main topic of quality of care issues inmaternity care. Although the Acute Physiology and Chronic HealthEvaluation II score is commonly used in the intensive care unit, itdoes not seem to be appropriate for pregnant women because itoverestimates their mortality rates. A high-dependency care unitsuits the needs for at least half of the obstetric patient populationin need of higher acuity care and will save considerable cost.

Conclusion: Emphasis on early detection of maternal problemsand prompt referral to tertiary centers with intensive care unitfacilities to provide optimum care of the circulation, blood pres-sure, and respiration at an early stage could minimize the prev-alence of multiple organ failure and mortality in critically illobstetric patients. (Crit Care Med 2006; 34[Suppl.]:S208–S214)

KEY WORDS: pregnancy; preeclampsia; critical care; obstetrichemorrhage; maternal mortality

S208 Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

Table 1. Overview of literature

Author(Reference No.)

YearCity,

CountryTime

Span, yrs

No. (%)of

Deliveries

AdmissionDiagnosis,

%

AntepartumPostpartum,

%

MaternalDeath, n

(%)

PerinatalDeath,

%

ChronicMedical

Disease, %

ICUUtilization

Rate, %

Duration ofICU Stay,

Days

Mabie and Memphis, 3 200 (0.9) P � 46 NS 7 (3.5) NS NS 0.9 Mean, 2.5Sibai (2) TN, US H � 101990

Kilpatrick and San Francisco, 5 32 (0.4) P � 22 AP � 34 4 (12) NS NS NS Mean, 5.4Matthay (5) CA, US H � 161992

Collop and Charleston, 3.3 20 (0.3) P � 30 AP � 85 4 (20) 35 50 NS Mean, 8Sahn (4) SC, US H � NS PP � 151993

Monaco et al. Chapel Hill, 8 38 (0.3) P � 32 AP � 29 7 (18) 12 NS NS Mean, 5.4(6) 1993 NC, US H � 5 PP � 71

Umo-Etuk et al. London, UK 5 39 (0.6) P � 33 AP � 15 1 (0.4) NS 26 NS Median, 2(9) 1996 H � 33 PP � 85

Wheatley et al. London, UK 5 144 (0.9) P � 66 AP � 8 3 (4.3) 6 16 12 Median, 3(22) 1996 H � 19 PP � 92

Tang et al. Hong Kong, 8 49 (0.1) P � 14 NS 2 (4.1) 10 NS NS NS(11) 1997 China H � 53

Platteau et al. Durban, South 1 122 (NS) P � 66 NS 26 (21) NS NS 14 Mean, 4.4(10) 1997 Africa H � NS

Lapinsky et al. Toronto, 5 65 (0.3) P � 40 AP � 9 0 11 29 NS Mean, 2.9(7) 1997 Canada PP � 91

Bewley and London, UK 2 30 (0.5) P � 40 NS 2 (7) NS 28 NS Median, 2Creighton H � 47(46) 1997

Baskett and Halifax, 14 55 (0.1) P � 25 AP � 18 2 (3.6) NS NS NS Mean, 5Sternadel Canada H � 22 PP � 82(47) 1998

Mahutte et al. Montreal, 8 131 (0.3) P � 21 AP � 23 3 (2.3) NS 38 NS Mean, 2.5(48) 1999 Canada H � 26 PP � 78

Cohen et al. Petak Tiqva, 4 46 (0.2) P � 32 AP � 7 1 (0.5) 22 NS 2 Mean, 2(49) 2000 Israel H � 24 PP � 93

Ryan et al. Dublin, 4 123 (1.0) P � 45 AP � 15 0 8 NS NS Median, 3(24) 2000 Ireland H � 21 PP � 85

Panchal et al. Maryland, US 14 1023 (0.1) P � 37 NS 34 (3) NS 8.6 NS Median, 2(50) 2000 (Statewide) H � 29

Quah et al. Singapore, NS 239 (0.7) P � 50 NS 3 (1.3) 4 NS NS Median, 2(51) 2001 Singapore H � 24

Afessa et al. Jacksonville, 8 74 (NS) P � 45 AP � 42 2 (3) 11 50 NS Median, 2.5(52) 2001 FL, US H � NS PP � 58

Loverro et al. Bari, Italy 11 41 (0.2) P � 76 AP � 0 2 (4.9) 111 17 NS Mean, 5(53) 2001 H � 15 PP � 100

Hazelgrove et al. Southern UK 3 210 (0.2) P � 40 AP � 19 7 (3.3) 20 NS 2 Median, 1(19) 2001 H � 33 PP � 81

Olarra et al. Madrid, Spain 7 149 (0.3) P � 50 NS 11 (7.5) 13 45 NS NS(54) 2002 H � NS

Heinonen et al. Kuopio, 6.5 22 (0.1) P � 32 AP � 3 1 (5) NS 4.5 NS Median, 6(55) 2002 Finland PP � 19

Gilbert et al. New Brunswick, 8 233 (0.5) P � 46 AP � 28 8 (3) NS 42 NS Mean, 4(17) 2003 NJ, US H � 24 PP � 72

Cheng and Singapore, 5 43 (0.3) P � 35 AP � 6 2 (5) NS NS 1 Median, 3Raman (56) Singapore H � 40 PP � 332003

Demirkiran et al. Istanbul, 5 125 (0.9) P � 74 AP � 1 13 (10) NS NS 3 Mean, 4(57) 2003 Turkey H � 11 PP � 124

Zeeman et al. Dallas, 2 483 (1.7) P � 42 AP � 20 1 (0.2) NS 13 NS Median, 1(3) 2003 TX, US PP � 80

Okafor and Enugu, 6 18 (0.3) P � 50 AP � 1 6 (33) NS NS 2 Median, 2.5Aniebue (58) Nigeria PP � 172004

Karnad et al. Mumbai, 5 453 (0.5) P � 56 AP � 38 98 (22) 52 8 NS Median, 4(15) 2004 India H � 24 PP � 62

Anwari et al. Riyadh, Saudi 6 99 (0.2) P � 29 AP � 1 1 (1) NS 11 2 Mean, 2(59) 2004 Arabia H � 32 PP � 98

Mirghani et al. Abu Dhabi, 6 60 (0.3) P � 25 NS 2 (3.3) NS 32 2 Mean, 2(23) 2004 Arab Emirates H � 28

Selo-Ojeme et al. London, UK 10 33 (0.1) P � 39 AP � 10 1 (3) NS 15 1 Mean, 8(60) 2005 H � 36 PP � 90

ICU, intensive care unit; P, preeclampsia; H, hemorrhage; AP, antepartum; PP, postpartum; NS, not stated.

S209Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

was mechanical ventilation. Most of suchpatients underwent general anesthesia foremergency cesarean section. In those cases,considerable intraoperative blood loss andreplacement required short-term continu-ation of mechanical ventilation in the post-operative period. Prolonged ventilatory orinotropic support was only needed in a mi-nority of these patients.

As can be seen in Table 1, the mostcommon reasons for ICU admission werepreeclampsia-related complications andpostpartum hemorrhage. Causes of hemor-rhage included placenta previa, placentalabruption, uterine atony, and often second-ary, profound coagulopathies. Preeclamp-sia-related complications include eclamp-sia, intracerebral hemorrhage, pulmonaryedema, renal insufficiency, liver rupture,and placental abruption. In South Africa,pregnancy-related hypertensive disordersaccounted for 66.4% of ICU admissions(10), whereas in China, massive postpar-tum hemorrhage accounted for 53% (11).Perhaps this is a reflection of the preva-lence of these conditions in those parts ofthe world during pregnancy and delivery.Studies from India, Sri Lanka, and Brazilhave shown that rheumatic valvular heartdisease, cerebral venous thrombosis, ma-laria, and viral hepatitis are important rea-sons for ICU admission during pregnancyand also significant causes of maternalmortality (12–15).

The cited studies are interesting butalso have major limitations; data are usu-ally collected retrospectively and samplesizes are small. This makes it difficult todetermine the prognostic factors and dif-ferences between survivors and nonsurvi-vors. Second, when studies are performedin tertiary care centers, the conclusionsmay not apply to other patient popula-tions. Lastly, it is generally not possibleto identify preventable conditions thatmay have led to ICU admission.

Lack of ICU Scoring Systems forCritically Ill Obstetric Patients

Little is known regarding the ability toassess severity of illness and predict out-comes in obstetric patients. There are nomodels that are designed specifically foruse in obstetric patients. Several investi-gators have applied a variety of scoringtools derived from nonobstetric popula-tions, such as the Acute Physiology andChronic Health Evaluation (APACHE) II,Mortality Probability Model, and the Sim-plified Acute Physiology Score (SAPS) II,to obstetric populations in an attempt to

predict the probability of hospital mortal-ity, with conflicting results (7, 16–19).

APACHE and SAPS II tend to overes-timate mortality in obstetric patients (17,19, 20). This is partly because normalphysiologic variables in obstetric patientsare scored as abnormal, such as the lowercreatinine, blood urea nitrogen, and he-matocrit and the higher respiratory rateand heart rate of pregnancy. Anotherproblem with the current scores is thatfor obstetric patients with hypertensive dis-orders, tests of liver function and plate-let count are important in the assess-ment. These variables are not affected inavailable scoring systems for critical ill-ness. In general, one can conclude thatwhen obstetric patients are admitted formedical disorders, the predicted mortal-ity rate more often matches the observedmortality rate. On the other hand, in pa-tients with obstetric disorders, observedmortality rate is usually much lower thanthe predicted mortality rate.

Obstetric High-DependencyCare Unit

There are no detailed guidelines fromany specialty organization that describe aplan of care of critically ill obstetric pa-tients. Thus, it is reasonable to use theguidelines for intensive and intermediatecare from the American College of Criti-cal Care Medicine (21) as the basis forobstetric critical care, as summarized inTable 2. Owing to their special problems,combined with the frequent chronic shor-tage of costly ICU beds, some have sug-gested that obstetric patients should bereferred to specialized obstetric ICUs(22). This is not unreasonable becausetwo large descriptive reports showed thatabout half of all women were thought tohave illness that was sufficiently severe to

necessitate ICU admission, whereas theremaining half were thought to be suit-able for intermediate or high-dependencycare (19, 23). The intermediate or high-dependency care unit (HDU) may be ap-propriate for patients who are consciousand who have single-organ dysfunction.Patients requiring advanced respiratorysupport, two or more organ system sup-port, and those with chronic system in-sufficiency and requiring support foracute reversible failure of another organsystem are considered to need ICU admis-sion.

HDUs have not been assessed formallyfor obstetric patients; however, severalreferral centers have actually incorpo-rated these concepts. In this environ-ment, care is provided by one nurse totwo patients and usually supervised bymaternal–fetal medicine specialists andobstetric anesthesiologists rather than byintensive care specialists. Studies report-ing on the availability of an HDU reportlow ICU rates and HDU utilization ratesthat average 1% (2, 3, 24). This possiblysuggests a lower threshold for admissionto the HDU. Zeeman et al. (3) described a2-yr audit of the obstetric intermediatecare unit. General criteria for admissionto this unit that were in place during the2-yr audit are shown in Table 3. Criteriafor admission or transfer to a medical/surgical ICU include all women whoseconditions require mechanical ventilation.Other reasons included major trauma, neu-rosurgical injuries, or cardiothoracic in-juries. This intermediate care unit al-lowed for the continuation of care byobstetricians and resulted in fewer trans-fers to medical/surgical ICUs.

The advantages of an HDU within anobstetric setting are numerous, for in-stance, the concurrent availability of ex-pert obstetric care and critical care man-

Table 2. Summary of guidelines from the American College of Critical Care Medicine (1998) ofconditions or diseases that could qualify for intermediate care

Cardiac Rule out infarction, stable infarction, stable arrhythmias, mild-to-moderatecongestive heart failure, hypertensive urgency without end-organ damage

Pulmonary Stable patients for weaning and chronic ventilation, patients with potentialfor respiratory failure who are hemodynamically stable

Neurologic Stable central nervous system, neuromuscular, or neurosurgical conditionsthat require close monitoring

Drug overdose Hemodynamically stableGastrointestinal Stable bleeding, liver failure with stable vital signsEndocrine Diabetic ketoacidosis, thyrotoxicosis that requires frequent monitoringSurgical Postoperative major procedures or complications that require close

monitoringMiscellaneous Early sepsis, patients whose condition requires closely titrated intravenous

fluids, pregnant women with severe preeclampsia or other medicalproblem

S210 Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

agement. Antenatal patients admitted tothe HDU have the option of continuousfetal monitoring with on-hand expertisein its interpretation. The hazards ofemergency transport of the obstetric pa-tient differs from adult transport in sev-eral respects; there are two or more pa-tients and these women are particularlyat risk of improper positioning. The ob-vious advantages of keeping mother andinfant together combined with the im-proved continuity of antenatal and post-natal care are further benefits of criticalcare facilities within this setting.

Maternal Morbidity as Indicatorfor Quality of Care

Current maternal mortality rates indeveloped countries are extremely low,ranging from 2.4/100,000 in Canada to9.5/100,000 in Japan (25). Developingcountries report exponentially higherrates such as 340/100,000 in SouthAfrica, 440/100,000 deliveries in India, to2,300/100,000 in Rwanda (15, 26). Thisusually reflects the poor availability ofhealthcare services including prenatalcare. The two preventable factors thatadversely affect the maternal mortalityrate are inadequate utilization of prenatalservices and a delay of �24 hrs betweenonset of acute illness and ICU admission.

Maternal mortality rates are generallyused to evaluate the quality of maternalcare. Maternal mortality rates in North

American and European ICUs range from3% to 20%, whereas this may be doubledin less-developed regions of the world(Table 1). Heterogeneity of the patientpopulation and differences in disease se-verity may account for the differences inthe reported ICU mortality rates. In gen-eral, mortality rate is higher in patientswith medical disorders than in patientswith obstetric diseases. The question iswhether clinical guidelines and recom-mendations when based on unusualevents such as maternal death are ofmuch value. It is possible that informa-tion on severe acute maternal morbidityas indicated by “near-miss” cases and ob-stetric admission to the ICU are moremeaningful when assessing the quality ofcare (24). In fact, it has been shown that,after exclusion of maternal thromboem-bolism, there may be �100 near missesfor each direct maternal death (27). Thus,in western countries, where a dramaticdecrease in maternal mortality has beenobtained, puerperal morbidity has be-come the main topic of quality-of-careinvestigation, especially if severe enoughto require admission into the ICU.

Preeclampsia

Preeclampsia is a multiple-system dis-order that develops in previously normo-tensive women after 20 completed weeksof gestation. It is characterized by hyper-tension and proteinuria, both of whichresolve by the sixth postpartum week.Eclampsia is defined as the occurrenceof tonic–clonic convulsions in pregnantor puerperal women with preeclampsia.Other dramatic neurologic presentations,albeit uncommon, include blindness, al-tered state of consciousness, and coma.HELLP is a syndrome comprising hemo-lysis, elevated liver enzymes, and lowplatelets. It occurs in 4–12% of patientswith severe preeclampsia. Hypertensionmay or may not be present. Symptomsinclude vague symptoms of nausea andvomiting and right upper quadrant orepigastric pain.

Preeclampsia is relatively common,with the average rate of preeclampsia inthe United States estimated to be 26/1,000 deliveries (28). Recent estimates ofthe prevalence of eclampsia in the UnitedStates range from 0.6 to 3/1,000 livebirths (28–30). Along with hemorrhage,thromboembolism, and infection, pre-eclampsia is accountable for the world’slarge maternal mortality rates; approxi-mately 50,000–65,000 deaths are thought

to occur per year worldwide. In the UnitedStates, 23% of maternal deaths recorded in1997 were related to pregnancy hyperten-sion (1). The acute cerebral complicationsof (pre)eclampsia, such as intracranialhemorrhage or massive cerebral edema, ac-count for �75% of such fatalities, particu-larly in the presence of HELLP syndrome(31). Less well known is that (pre)eclampsiaalso accounts for nearly 50% of (mostlyclinically reversible) pregnancy-relatedischemic strokes (32). Improvement in an-tenatal and intensive care has reduced theprevalence of and death attributable toeclampsia in western countries during thepast decade. Modern maternal mortalityrates of �0.5% are now reported (33, 34).

Adequate oxygenation, urine output,and blood pressure are typical therapeu-tic goals in the management of criticallyill ICU patients, even more so in pre-eclampsia. Low cardiac filling pressures,high systemic vascular resistance, andlow cardiac output characterize this con-dition (1). This is thought to result fromaltered endothelial permeability, extrava-sation of plasma into the interstitium,and vasoconstriction.

Pulmonary Edema. A significantnumber of maternal deaths in preeclamp-sia are related to pulmonary edema. Therisk of the development of pulmonaryedema is aggravated by exogenous fluidadministered in the belief that thesewomen are at risk of renal failure. Renalfailure is, however, extremely unusual.Oliguria is fairly common after deliveryand does not require treatment becausethe large majority of postpartum womenwith preeclampsia will enter the diureticphase within the first 24–48 hrs postpar-tum. Fluid challenges are potentiallydangerous in preeclampsia because muchof the fluid will be lost from the vascula-ture into the interstitium. Only rarely isthe use of a pulmonary artery catheterrequired because persistent refractory ol-iguria in the presence of pulmonaryedema is extremely rare. The relationshipbetween central venous pressure and pul-monary artery occlusion pressure in se-vere preeclampsia is inconsistent. Hemo-dynamic monitoring using centralvenous pressure data in severe pre-eclampsia is unreliable and should not beused. Until a reliable noninvasive methodis available to measure left ventricularpreload, pulmonary artery occlusionpressure is the measurement of choicewhen invasive hemodynamic monitoringis necessary in patients with severe pre-eclampsia (35).

Table 3. Criteria for admission to the obstetricintermediate care unit or medical/surgical inten-sive care unit

Intermediate care unitObstetric complications

Severe preeclampsia or eclampsia orHELLP syndrome

Severe hemorrhage and/or coagulationdisorders

Complicated peripartum hysterectomyAcute fatty liver of pregnancySepsis

Surgical or anesthesia complicationsMedical or surgical disorders

Diabetic ketoacidosisThyrotoxicosisHemofiltration/plasmapheresisSevere asthma or pneumonia; complicated

cholecystitis, pancreatitis, or appendicitisMedical-surgical intensive care unit

Mechanical ventilationInotropic drugsLife-threatening arrhythmiaComa

HELLP is a syndrome comprising hemolysis,elevated liver enzymes, and low platelets.

S211Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

Hypertension. Aggressive blood pres-sure control is an important end point inthe treatment of these young patients inwhom acute increases of blood pressuremay pose significant risk. The aim is toprevent cerebral hemorrhage and hyper-tensive encephalopathy. The blood pres-sure in preeclampsia is variable and canfluctuate on a minute-to-minute basis.Blood pressure should be measured man-ually using standard aneroid sphygmo-manometers because automated bloodpressure monitors are notoriously unre-liable in women with (pre)eclampsia (36).The most commonly used antihyperten-sive agents are hydralazine, labetalol, andnifedipine. Therapy can be given by con-tinuous or bolus infusion, which ideallyrequires direct measurement of intraar-terial blood pressure to rapidly achieveend points while avoiding excessive de-crease in blood pressure.

Eclampsia. Convulsions occur antena-tally in 38%, intrapartum in 18% andpostpartum in the remaining 44%, usu-ally in the first 24–48 hrs after delivery.Although prodromal symptoms, mainlyheadache and right upper quadrant pain,are usually present in women who developeclampsia, about 20% might convulse un-expectantly, with a relatively normal bloodpressure and no specific signs for the de-velopment of eclampsia (28).

Eclampsia seems rarely associatedwith persistent, clinically recognizable,neurologic morbidity, and epilepsy is nota recognized long-term complication.However, intracerebral hemorrhage is awell-known complication of eclampsia,frequently leading to maternal death ormajor permanent disability. In general,therapy should be directed toward lower-ing blood pressure so as to limit the fur-ther development of vasogenic cerebraledema and subsequent ischemia.

Magnesium sulfate (MgSO4) is the an-ticonvulsant drug of choice because it iseffective for prevention and treatment ofseizures (Table 4) (37). In cases of recur-rent seizure, it is safe to give a secondbolus of 2 g of MgSO4. Polypharmacy is tobe avoided to treat seizures because thisdramatically increases the risk of respira-tory arrest. Only when seizures continue,despite administration of a second bolus,should diazepam or thiopental be admin-istered intravenously. Intubation thenbecomes necessary to protect the airwayand ensure adequate oxygenation. Fur-ther seizure activity should be managedby ventilation and muscle relaxation. Oneof the main concerns that many have is

the fear of magnesium toxicity. However,when using the cited dosage regimen inthe absence of renal insufficiency, ther-apy can be monitored safely by measure-ment of the patellar reflexes and respira-tory rate. In case of overdose, first ensureadequate ventilation, then 1 mg of 10%calcium gluconate can be given in a 10-min period.

On the basis of pathology and imagingfindings and the similarities in clinicalpresentation, attention has been directedto hypertensive encephalopathy as amodel for the cerebrovascular abnormal-ities in eclampsia (38). Strictly speaking,during a period of hypertensive encepha-lopathy, a relatively acute and excessiveintravascular pressure increase causesforced dilation of intrinsic myogenic toneof cerebral arteries. This decreases cere-brovascular resistance and increases pres-sure on the microcirculation, thereby caus-ing vasogenic edema formation. In theclinical setting, this phenomenon has re-cently been coined reversible posteriorleukoencephalopathy syndrome (39). En-dothelial dysfunction is generally consid-ered to play a key role in the clinicalmanifestation of reversible posterior leu-koencephalopathy syndrome and in pre-eclampsia.

Neuroimaging studies could generallybe limited to those women who have ad-ditional focal neurologic signs, prolongedcoma, atypical or recurrent convulsions,and those who have a prolonged return tocomplete recovery after delivery. In suchwomen, hemorrhage or other serious ab-normalities such as sinus thrombosis mustbe excluded. Magnetic resonance imagingis well known for its far more superiorsoft-tissue contrast and multiplanar resolu-tion compared with computed tomogra-phy. This technique is therefore extremelyeffective for the diagnosis of hemorrhageand ischemia or cerebral edema in women

with (pre)eclampsia. Whereas computed to-mography is often reported as normal ineclampsia, fluid-attenuated inversion re-covery (FLAIR) magnetic resonance imag-ing demonstrates transient T2 lesions inthe (sub)cortical regions of the parietoc-cipital lobes, suggestive of cerebral edema.Occasional involvement of basal ganglia orbrainstem is also reported (40). These hy-perintense lesions are typically thought toresolve without long-term sequelae. Stud-ies using additional diffusion-weighted im-aging sequences showed that the origin ofbrain edema in eclampsia is primarily va-sogenic but, less commonly, may be asso-ciated with ischemic/cytotoxic changes. Itseems that 20 –25% of women witheclampsia demonstrate minute lesions con-sistent with cerebral infarction on later fol-low-up, although such women seemedasymptomatic (40). The long-term signifi-cance of this is currently being investi-gated.

Hemorrhage

In the United States and other indus-trialized nations, massive obstetric hem-orrhage has generally ranked among thetop three causes of maternal death, de-spite modern improvements in obstetricpractice and transfusion services. Massiveobstetric hemorrhage usually occurs in-trapartum or within the first hour afterdelivery. Massive obstetric hemorrhage ismost commonly due to uterine atony,which complicates 5% of deliveries andresults in excessive blood loss when ade-quate myometrial contraction fails to oc-cur after placental expulsion.

The main factor in its management isthe restoration of the circulating blood vol-ume. A delay in the correction of hypovo-lemia, in diagnosis and treatment of im-paired coagulation, and in the surgicalcontrol of bleeding are the avoidable factorsin most maternal mortality cases caused byhemorrhage. Uterotonic agents, such asoxytocin, are used in the management ofuterine atony (41). This synthetic nonapep-tide is a first-line agent because of the pau-city of side effects and the absence of con-traindications. Methylergonovine, an ergotalkaloid, is used as a second-line uterotonicagent in the setting of massive obstetrichemorrhage due to atony. Hypertension isan absolute contraindication.

Injectable prostaglandins may also beused when oxytocin fails. ProstaglandinE2 and prostaglandin F2 stimulate myo-metrial contractions and have been usedfor refractory hemorrhage due to uterine

Table 4. Magnesium sulfate for prevention andtreatment of eclampsia

1. Loading dose: 4–6 g diluted in 100 mL ofintravenous fluid administered during15–20 mins (give in 5 mins instead whenactively seizing)

2. Maintenance dose of 1–2 g/hr in 100 mL ofintravenous maintenance infusion

3. Serum levels should be maintained betweena. 4–7 mEq/Lb. 4.8–8.4 mg/dLc. 2–4 mmol/L

4. Discontinue infusion 24 hrs after delivery orlast convulsion

S212 Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

atony. Such prostaglandins and their sev-eral analogs are used either intramuscu-larly or intravenously. A major contrain-dication to their use is asthma, and thereis a considerable risk for myocardial in-farction, when used in the presence ofsevere hypovolemia, due to coronary ar-tery spasm.

Recombinant activated factor VII(NovoSeven, Novo Nordisk A/S, Bagsvaerd,Denmark) has recently shown to be an ad-junctive hemostatic measure for the treat-ment of severe obstetric hemorrhage (42).Based on the mechanisms of action, cir-culating factor VII is active after it bindsto tissue factor, which is exposed at sitesof vessel injury. This complex initiatescoagulation on activated platelet surfacesadhering to the site of injury and result-ing into formation of a localized fibrinclot. The drug can be administered inobstetric cases with life-threatening hem-orrhage, even in the presence of dissem-inated intravascular coagulation–like co-agulopathy. A dose of 90 –200 �g/kgseems to be appropriate. Controlled tri-als, however, have not been performed.

The use of a hydrostatic balloon hasbeen advocated as an alternative to uter-ine packing for controlling hemorrhagedue to uterine atony (43). The inflatedRusch balloon can conform to the con-tour of the uterine cavity and provides aneffective tamponade. Life-threateninghemorrhage can also be treated by arte-rial embolization (44). In cases of con-tinuing hemorrhage, a variety of surgicaltechniques can be used to avoid a hyster-ectomy, such as bilateral uterine arteryligation or internal iliac artery ligation.More recently, the use of a B-Lynch bracesuture, which compresses the uteruswithout compromise of the major vessels,has been advocated (45). In general, pooroutcomes stem from a delay in perform-ing any of the above measures or a firmhesitation in performing a lifesaving hys-terectomy.

Who Cares for the Critically IllObstetric Patient?

The high variability of mortality ratesof critically ill obstetric patients corre-lates not only with the severity of under-lying disease but also with clinical recog-nition of the unique needs of this patientpopulation. Critically ill obstetric pa-tients present an interesting challenge interms of medical management and oftenrequire the input of several specialties.Most centers report that obstetric patients

are treated by critical care staff physicians,often represented by anesthetists or in-tensivists. Monitoring is often invasiveand requires special skills. Most obstetri-cians do not see sufficient cases to ac-quire and maintain these skills. Therecan be little doubt that intensivists in anICU best treat these patients. Likewise, itis crucial that someone who is knowl-edgeable in maternal and fetal physiologybe included in the treatment team, par-ticularly in the care for the preeclampticwoman. Care must include the consider-ation of pregnancy-induced physiologicchanges, normal laboratory alterations,and continued fetal well-being if antepar-tum. This presents unique medical andethical dilemmas with which obstetri-cians are more familiar. Ideally, a work-ing group that consists of members fromboth disciplines routinely reviews all ad-missions. The goal of this group is toensure cohesive coordinated care.

Conclusions

Hypertension-related complicationsand massive obstetric hemorrhage due touterine atony in previously low-riskwomen are the leading causes for admis-sion of obstetric patients to the ICU. Theadmission rate to intensive care and theproblems faced by critically ill parturientsmay be reduced by improving the man-agement of hypertensive disease duringpregnancy and by reducing the preva-lence of hemorrhagic complications byemphasis on early detection and anticipa-tion. When complications arise early, in-tervention and treatment on a multidis-ciplinary basis, which may involve ICUadmission for ventilatory support, inva-sive monitoring, and vasoactive drug in-fusions, can alleviate progression of or-gan dysfunction and improve prognosis.

Institutional capabilities and the fre-quency and acuity of serious obstetriccomplications largely will be the mainfactors to determine the need for criticalcare facilities. An intermediate care unitshould fulfil the needs of most tertiarycare centers. Academic centers will ben-efit because such units are important forresident and fellow education and experi-ence. Smaller hospitals may not be ableto fulfill the requirements for an inter-mediate care unit (21), or they may notencounter enough critically ill women tomaintain contemporaneous skills. Impli-cations regarding individual hospital vol-ume of deliveries, as the volume relatesto the care of the critically ill obstetric

patients, is yet to be determined and re-quires future investigation.

Maternal death has become an ex-tremely rare event in developed coun-tries, which weakens the value of suchrates as a quality-assurance indicator formaternity care. There is growing interestin the use of major morbidity, so-callednear miss, as an indicator of the quality ofhospital-based obstetric care. The need totransfer to the ICU is used as an indicatorof illness severity. Scoring systems suchas the APACHE II or the SAPS score ac-curately predict hospital mortality amongobstetric patients admitted to the ICU formedical reasons but perform poorly inpredicting deaths from patients admittedpurely for obstetric reasons. Ideally, fu-ture studies should try to estimate thepredictor factors of severe obstetric mor-bidity to improve prenatal care, perinatalmanagement, and anesthetic procedures.With reliable predictor factors, the num-ber of obstetric patients who require crit-ical care would be reduced and the ratesof maternal and fetal morbidity and mor-tality lowered.

REFERENCES

1. Cunningham FG, Leveno KJ, Bloom SL, etal: Williams Obstetrics. 22nd Edition. NewYork, McGraw-Hill, 2005

2. Mabie WC, Sibai BM: Treatment in an obstet-ric intensive care unit. Am J Obstet Gynecol1990; 162:1–4

3. Zeeman GG, Wendel GD, Cunningham FG: ABlueprint for obstetric critical care. Am JObstet Gynecol 2003; 188:532–536

4. Collop NA, Sahn SA: Critical illness in preg-nancy: Analysis of 20 patients admitted to amedical intensive care unit. Chest 1993; 103:1548–1552

5. Kilpatrick SJ, Matthay MA: Obstetric patientsrequiring critical care. Chest 1992; 101:1407–1412

6. Monaco TJ, Spielman RJ, Katz VL: Pregnantpatients in the intensive care unit. SouthMed J 1993; 86:414–417

7. Lapinsky SE, Kruczynski K, Seaward GR, etal: Critical care management of the obstetricpatient. Can J Anaesth 1997; 544:325–329

8. Ancel PY, Bouvier-Colle MH, Breart G, et al:Risk factors for maternal condition at admis-sion to an intensive care unit: Does healthcare organisation play a role? Study group ofthe maternal morbidity. J Perinat Med 1998;26:354–364

9. Umo-Etuk J, Lumley J, Holdcroft A: Criti-cally ill parturient women and admission tointensive care: A 5-year review. Int J ObstetAnesth 1996; 5:79–84

10. Platteau P, Engelhardt T, Moodley J, et al:Obstetric and gynaecological patients in an

S213Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)

intensive care unit: A 1 year review. TropDoct 1997; 27:202–206

11. Tang LC, Kwok AC, Wong AY, et al: Criticalcare in obstetrical patients: An eight-year re-view. Chin Med J (Engl) 1997; 110:936–941

12. Dao B, Rouamba A, Ouedraogo D, et al:Transfer of obstetric patients in a pregnantand postpartum condition to an intensivecare unit: Eighty-two cases in Burkina Faso.Gynecol Obstet Fertil 2003; 31:123–126

13. Dias de Souza JP, Duarte G, Basile-Filho A:Near-miss maternal mortality in developingcountries. Eur J Obstet Gynecol Reprod Biol2002; 104:80

14. Wagaarachi PT, Fernando I: The impact of anintensive care unit on maternal mortality.Int J Gynecol Obstet 2001; 74:199–201

15. Karnad DR, Lapsia V, Krishnan A, et al: Prog-nostic factors in obstetric patients admittedto an Indian intensive care unit. Crit CareMed 2004; 32:1294–1299

16. Lewinsohn G, Herman A, Leonov Y, et al:Critically ill obstetrical patients: Outcomeand predictability. Crit Care Med 1994; 22:1412–1414

17. Gilbert TT, Smulian JC, Martin AA, et al:Obstetric admissions to the intensive careunit: Outcomes and severity of illness. ObstetGynecol 2003; 102:897–903

18. Bhagwanjee S, Paruk F, Moodley J, et al:Intensive care unit morbidity and mortalityfrom eclampsia: An evaluation of the AcutePhysiology and Chronic Health Evaluation IIscore and the Glasgow Coma Scale score.Crit Care Med 2000; 28:120–124

19. Hazelgrove JF, Price C, Pappachan VJ, et al:Multicenter study of obstetric admissions to14 intensive care units in southern England.Crit Care Med 2001; 29:770–775

20. Scarpinato L: Critical illness in pregnancy.Chest 1994; 105:1915–1916

21. American College of Critical Care Medicine andthe Society of Critical Care Medicine: Guide-lines on admission and discharge for adult in-termediate care units: Guidelines/Practice Pa-rameters Committee of the American Collegeof Critical Care Medicine and the Society ofCritical Care Medicine. Crit Care Med 1998;26:607–610

22. Wheatley E, Farkas A, Watson D: Obstetricadmissions to an intensive therapy unit. Int JObstet Anesth 1996; 5:221–224

23. Mirghani HM, Hamed M, Ezimokhai M, et al:Pregnancy-related admissions to the inten-sive care unit. Int J Obstet Anesth 2004;13:82–85

24. Ryan M, Hamilton V, Bowen M, et al: Therole of a high-dependency unit in a regionalobstetric hospital. Anaesthesia 2000; 55:1155–1158

25. Nagaya K, Fetters MD, Ishikawa M, et al:Causes of maternal mortality in Japan. JAMA2000; 283:2661–2667

26. Maine D, Chavkin W: Maternal mortality:Global similarities and differences. J Am MedWomens Assoc 2002; 57:127–130

27. Waterstone M, Bewley S, Wolfe C: Incidenceand predictors of severe obstetric morbidity:Case-control study. BMJ 2001; 322:1089–1093

28. Saftlas AF, Olson DR, Franks AL, et al: Epi-demiology of preeclampsia and eclampsia inthe United States, 1979–1986. Am J ObstetGynecol 1990; 163:460–465

29. Mattar F, Sibai BM: Eclampsia: VIII. Riskfactors for maternal morbidity. Am J ObstetGynecol 2000; 182:307–312

30. Abi-Said D, Annegers JF, Combs-Cantrell D,et al: Case-control study of the risk factorsfor eclampsia. Am J Epidemiol 1995; 142:437–441

31. Okanloma KA, Moodley J: Neurological com-plications associated with the preeclampsia/eclampsia syndrome. Int J Gynaecol Obstet2000; 71:223–225

32. Kittner SJ, Stern BJ, Feeser BR, et al: Preg-nancy and the risk of stroke. N Engl J Med1996; 12:335:768–774

33. Sibai BM: Eclampsia. VI: Maternal-perinataloutcome in 254 consecutive cases. Am J Ob-stet Gynecol 1990; 163:1049–1054

34. Pritchard JA, Cunningham FG, Pritchard SA:The Parkland Memorial Hospital protocol fortreatment of eclampsia: Evaluation of 245cases. Am J Obstet Gynecol 1984; 148:951–963

35. Bolte AC, Dekker GA, van Eyck J, et al: Lackof agreement between central venous pres-sure and pulmonary capillary wedge pressurein preeclampsia. Hypertens Pregnancy 2000;19:261–271

36. Shennan AH, Halligan AW: Measuring bloodpressure in normal and hypertensive preg-nancy. Baillieres Best Pract Res Clin ObstetGynaecol 1999; 13:1–26

37. Altman D, Carroli G, Duley L, et al: Dowomen with pre-eclampsia, and their babies,benefit from magnesium sulphate? The Mag-pie Trial: A randomised placebo-controlledtrial. Lancet 2002; 359:1877–1890

38. Schwartz RB, Feske SK, Polak JF, et al: Pre-eclampsia-eclampsia: Clinical and neurora-diographic correlates and insights into thepathogenesis of hypertensive encephalopa-thy. Radiology 2000; 217:371–376

39. Hinchey J, Chaves C, Appignani B, et al: Areversible posterior leukoencephalopathysyndrome. N Engl J Med 1996; 334:494–500

40. Zeeman GG, Fleckenstein JL, Twickler DM,et al: Cerebral infarction in eclampsia. Am JObstet Gynecol 2004; 190:714–720

41. Dildy GA: Postpartum hemorrhage: New man-agement options. Clin Obstet Gynecol 2002;45:2:330–344

42. Segal S, Shemesh IY, Blumenthal R, et al:Treatment of obstetric hemorrhage with re-combinant activated factor VII (rFVIIa). ArchGynecol Obstet 2003; 268:266–267

43. Johanson R, Kumar M, Obhrai M, et al: Man-agement of massive postpartum haemorrhage:Use of a hydrostatic balloon catheter to avoidlaparotomy. BJOG 2001; 108:420–422

44. Pelage JP, Le Dref O, Mateo J, et al: Life-threatening primary postpartum hemor-

rhage: Treatment with emergency selectivearterial embolization. Radiology 1998; 208:359–362

45. B-Lynch C, Coker A, Lawal AH, et al: TheB-Lynch surgical technique for the control ofmassive postpartum haemorrhage: An alter-native to hysterectomy? Five cases reported.Br J Obstet Gynaecol 1997; 104:372–375

46. Bewley S, Creighton S: “Near-miss” obstetricenquiry. J Obstet Gynaecol 1997; 17:26–29

47. Baskett RF, Sternadel J: Maternal intensivecare and near-miss mortality in obstetric.Br J Obstet Gynaecol 1998; 105:981–984

48. Mahutte NG, Murphy-Kaulbeck L, Le Q, et al:Obstetric admissions to the intensive careunit. Obstet Gynecol 1999; 94:263–266

49. Cohen J, Singer P, Kogan A, et al: Course andoutcome of obstetric patients in a generalintensive care unit. Acta Obstet GynecolScand 2000; 79:846–850

50. Panchal S, Arria A, Harris AP: Intensive careutilization during hospital admission for de-livery: Prevalence, risk factors, and outcomesin a statewide population. Anesthesiology2000; 92:1537–1544

51. Quah TC, Chiu JW, Tan KH, et al: Obstetricadmissions to the intensive therapy unit of atertiary care institution. Ann Acad Med Sin-gapore 2001; 30:250–253

52. Afessa B, Green B, Delke I, et al: Systemicinflammatory response syndrome, organ fail-ure, and outcome in critically ill obstetricpatients treated in an ICU. Chest 2001; 120:1271–1277

53. Loverro G, Pansini V, Greco P, et al: Indica-tions and outcome for intensive care unitadmission during puerperium. Arch GynecolObstet 2001; 265:195–198

54. Olarra J, Longarela AM, Suarez L, et al: Crit-ically ill obstetric patients treated in an ICU.Chest 2002; 121:2077

55. Heinonen S, Tyrväinen, Saarikoski S, et al:Need for maternal critical care in obstetrics:A population-based analysis. Int J ObstetAnesth 2002; 11:260–264

56. Cheng C, Raman S: Intensive care use bycritically ill obstetric patients: A five-year re-view. Int J Obstet Anesth 2003; 12:89–92

57. Demirkiran O, Dikmen Y, Utku T, et al: Crit-ically ill obstetric patients in the intensivecare unit. Int J Obstet Anesth 2003; 12:266–270

58. Okafor UV, Aniebue U: Admission pattern andoutcome in critical care obstetric patients.Int J Obstet Anesth 2004; 13:164–166

59. Anwari JS, Butt AA, Al-Dar MA: Obstetricadmissions to the intensive care unit. SaudiMed J 2004; 25:1394–1399

60. Selo-Ojeme DO, Omosaiye M, Battacharjee P,et al: Risk factors for obstetric admissions tothe intensive care unit in a tertiary hospital:A case-control study. Arch Gynecol Obstet2005; 272:207–210

S214 Crit Care Med 2006 Vol. 34, No. 9 (Suppl.)