Acta Obstet Gynecol Scand 2000; 79: 846–850 Copyright C Acta Obstet Gynecol Scand 2000

Printed in Denmark ¡ All rights reservedActa Obstetricia et

Gynecologica ScandinavicaISSN 0001-6349

ORIGINAL ARTICLE

Course and outcome of obstetric patients in ageneral intensive care unitJONATHAN COHEN1, PIERRE SINGER1, ALEX KOGAN1, MOSHE HOD2 AND JACOB BAR2

From the 1Department of General Intensive Care and the 2Perinatal Division, Department of Gynecology and Obstetrics, aWHO Collaborating Center on Perinatal Care, Rabin Medical Center, Beilinson Campus, Petah Tiqva, Israel

Acta Obstet Gynecol Scand 2000; 79: 846–850. C Acta Obstet Gynecol Scand 2000

Background. To characterize the course, interventions required to achieve predetermined end-points and outcome of obstetric patients admitted to a general intensive care unit.Methods. A retrospective case series study was performed including all pregnant patientsadmitted to an 8-bed general intensive care unit at a tertiary care university-affiliated hospitalover a 4-year period. All patients referred by the obstetricians were admitted. Patients weredivided into two groups: group 1, (nΩ19) those requiring mechanical ventilatory support andgroup 2, (nΩ27) those requiring intensive monitoring. Data collected included demographics,reason for admission, admission diagnosis, Acute Physiology and Chronic Health Evaluation(APACHE II) and Therapeutic Intervention Scoring System (TISS) scores, intensive care unitcourse, types of interventions used and outcome. End-points of therapy included systolicblood pressure 110–150 mmHg, urine output Ø1 cc/kg/h and oxygen saturation .95%.Results. Over the study period, 46 obstetric patients were admitted to the intensive care unit,representing 0.2% of all deliveries and an intensive care unit utilization rate of 2.3%. Com-monest admission diagnoses were pregnancy-induced hypertension and hemorrhage. Reasonfor admission was mechanical ventilation in 41% while 59% were admitted for monitoring.Median length of stay was 25∫80.9 (mean 48.8) hours. The median APACHE II score was6∫3.9 (mean 7.24) and the TISS score was .20 in both groups. Only one patient died (mor-tality rate 2.3%).Conclusion. Despite a short length of stay and low APACHE score, the high TISS score inobstetric patients admitted for both ventilation and monitoring suggests that these patientsrequire a level of intervention and care typically provided by a general intensive care unit.

Key words: complications; general intensive care unit; obstetric

Submitted 30 August, 1999Accepted 7 February, 2000

Obstetric patients are generally young and healthy.However, the potential for dramatic and even cata-strophic complications is real. These may be re-lated to the pregnancy itself, aggravation of a pre-existing illness, or complications of the frequentprocedures these patients undergo. Obstetric pa-tients with serious complications are usually re-

Abbreviations:ICU: intensive care unit; SBP: systolic blood pressure;APACHE: acute physiologic and chronic health evaluation;TISS: therapeutic intervention scoring system; HELLP: hemo-lysis, elevated liver enzymes, low platelets.

C Acta Obstet Gynecol Scand 79 (2000)

ferred to intensive care units (ICUs) for ventilationand critical management. Owing to their specialproblems, combined with the frequent chronicshortage of costly ICU beds, some researchershave suggested that these patients should be re-ferred to specialized obstetric ICUs (1).

There is another larger group of obstetric pa-tients who require intensive monitoring in order toprevent serious consequences. This monitoring isoften invasive and requires special skills. However,most obstetricians do not see sufficient cases to ac-quire and maintain these skills. Additional prob-lems are the need for dedicated staff and facilities

Obstetric patients in General ICU 847

and lack of time and space in most large obstetricdepartments.

As there is no dedicated area in our obstetric de-partment for the treatment of patients requiringeither mechanical ventilation or intensive and invas-ive monitoring, all such patients are admitted to ourgeneral ICU. This allows us to acquire much experi-ence with these patients. The aim of the presentstudy is to characterize their course and outcome.

Material and methods

The present study was performed at the RabinMedical Center, Beilinson Campus, a major, 850-bed, tertiary care, university-affiliated hospital. Wereviewed the files of all patients admitted to ourseven-bed multidisciplinary general ICU betweenJanuary 1994 and July 1998 who were pregnant(estimated gestational age at least 20 weeks) or inthe immediate (up to 2 weeks) postpartum period.All patients who were referred by their obstetricianwere admitted, and no patient was denied admis-sion because of lack of a bed, even if this meantdischarging a current patient. Approval of thestudy by the local Helsinki Committee was deemedunnecessary because we used a retrospective case-study design.

Treatment decisions were made by the ICU staff,comprised of full-time intensive care specialists,and obstetricians were consulted when necessary.Decisions about delivery were made only by ob-stetricians. End-points of therapy were systolicblood pressure (SBP) 110–150 mmHg, urine out-put Ø1 cc/kg/h, and oxygen saturation .95%. Pa-tients were discharged according to standard cri-teria.

The following data were obtained from the medi-cal records: demographic data; admission diag-nosis; reason for admission to the ICU; scores onthe Acute Physiology and Chronic Health Evalu-ation (APACHE II) (2) and the Therapeutic Inter-vention Scoring System (TISS) (3) at 24 hours of ad-mission; duration of mechanical ventilation (whennecessary); duration of ICU stay; admission hemo-globin and base excess values; main therapeutic in-terventions; maternal and fetal outcome. Resultsare given as the median∫standard deviation(mean). Student’s t-test was used to detect signifi-cant differences between the two groups while theMann-Whitney’s test was used to detect significantdifferences between duration of stay; a p value of,0.05, two-tailed test, was considered significant.

Results

Demographic data are shown in Table I. All weresingleton pregnancies. Forty-three patients (93%)

C Acta Obstet Gynecol Scand 79 (2000)

were admitted in the immediate postpartumperiod; of these, 35 (76%) were delivered by cesar-ean section. Of the three patients who were admit-ted prior to delivery, one had an ovarian abscesswhich required drainage, one had acute cholecys-titis which was later operated on, and one hadpneumonia. Two of these patients underwent sub-sequent cesarean section due to fetal distress.

Admission diagnoses and reason for critical careare shown in Table II. The most common admis-sion diagnoses were pregnancy-induced hyperten-

Table I. Demographic data

Obstetric admissions to ICU 46 (0.2% of all deliveries)Total deliveries over study period 19,474Total admissions to ICU over study period 1,985ICU utilization rate 2.3%Median patient age 30∫5.54 years (30.88)Median gestational age 32∫6 weeks (33.1)Patients after cesarean section 35 (76%)Predelivery admissions 3 (6.5%)

Results expressed as median∫s.d. (mean).

Table II. Admission diagnoses by reason for critical care

WholeAdmission diagnosis cohort Group I Group II

Hypertensive disease in pregnancy 15 3 12- preeclampsia/eclampsia 11 1 10- HELLP syndrome 4 2 2

Uterine hemorrhages 11 8 3- antepartum 2 1 1- intraoperative 5 4 1- postpartum 4 3 1

Surgical after-care 6 4 2- renal transplant patient 1 1- heroin abuse 1 1- bladder tear 1 1- septic abortion 1 1- drainage of ovarian abscess 1 1- acute cholecystitis 1 1

Medical 6 1 5- pneumonia 3 3- acute MI post C/S 1 1- pulmonary embolism 1 1- coumadin bleed 1 1

Pulmonary edema 4 2 2- sepsis 3- cause unknown 1

Post-anesthetic complication 4- confusion 2 1 3- hypotension 1 2- aspiration 1 1 1

Total (%) 46 19 (41%) 27 (59%)

HELLPΩhemolysis, elevated liver enzymes, low platelet count.C/SΩcesarean section.Group I – patients requiring ventilation.Group II – patients requiring invasive or intensive monitoring.

848 J. Cohen et al.

Table III. ICU course for whole cohort

ICU stay 25∫80.89 hours (48.82),24 hours 20 patients (44%)24–48 hours 17 patients (37%).48 hours 9 patients (19%)

No. ventilated (%) 19 patients (41%)

Duration of ventilation 12∫57.04 hours (32.18),24 hours 13 patients (68%)

APACHE II score (nΩ26) 6∫3.94 (7.24)

TISS score (nΩ24) 24∫6.09 (25.28),10 points 0 patients10–19 points 5 patients20–29 points 15 patients30–39 points 4 patients.40 points 0 patients

Admission laboratory results- base excess ª4.9∫3.33 (ª5.06)- hemoglobin 9.5∫1.56 g/dl (9.8)

Outcome- survived 45 (97.8%)- died 1 (2.2%)- neonatal mortality 10 (22.6%)

Results expressed as median∫s.d. (mean).

sion (32%) and excessive uterine hemorrhage(24%). Nineteen patients (41%) required mechan-ical ventilation. The reason for ventilation wascontinuing hemodynamic instability followingsurgery either because of, or complicated by, ex-cessive uterine hemorrhage (eight patients); con-tinuing severe hypertension following emergencysurgery for pre-eclampsia and HELLP syndrome(three patients); anticipated difficulty in weaningfollowing emergency surgery for septic abortion,ovarian abscess, acute cholecystitis and bladdertear (four patients); acute myocardial infarctionfollowing emergency cesarean section (one pa-tient); acute pulmonary edema (two patients) andaspiration (one patient). For patients admitted formonitoring, an indwelling intraarterial catheter

Table IV. ICU course by reason for critical care

Group I Group IIParameter (nΩ19) (nΩ27) p value

Duration of stay (h) 39∫113.85 (74.58) 22∫12.41 (26.0) 0.004APACHE II score 9∫5.02 (9.92) 3∫2.02 (3.66) ,0.001TISS score 26∫5.30 (29.23) 20.0∫3.80 (20.09) ,0.001Base excess ª6.4∫3.63 (ª6.9) ª3.4∫2.76 (ª3.88) 0.035Hemoglobin (g/dl) 9.2∫1.8 (9.4) 9.8∫1.6 (10.22) 0.242Mortality (n) 0/19 1/27

Results expressed as median∫s.d. (mean).APACHE – Acute Physiological and Chronic Health Evaluation.TISS – Therapeutic Intervention Scoring System.Group I – patients requiring mechanical ventilation.Group II – patients requiring invasive and/or intensive monitoring.

C Acta Obstet Gynecol Scand 79 (2000)

was inserted in those requiring continuous arterialpressure monitoring and a subclavian vein catheterin those requiring central venous pressure monitor-ing. The ICU course is shown in Table III. Themedian duration of ICU stay was 25∫80.9 (48.82)hours, and the duration of ventilation was12∫57.04 (32.18) hours. For the 26 patients admit-ted for .24 hours, the median APACHE II scorewas 6∫3.9 (7.24), and the median TISS score was24∫6.1 (25.28). The TISS score was not availablein two patients for lack of complete data.

Analysis of outcome yielded one maternaldeath. This occurred in a 22-year-old primigravidaadmitted to the ICU for blood pressure monitor-ing following emergency cesarean section for ec-lampsia, seizures and fetal distress. She was dis-charged after 14 hours and readmitted the next dayafter having another seizure. Soon after admissionshe had a sudden cardiac arrest and died. The neo-natal mortality rate was 22%, mainly because ofextreme prematurity.

The patients were divided into two groups byreason for critical care: ventilation (Group I) orintensive or invasive monitoring (Group II) andcompared for admission diagnosis (Table I) andICU course (Table IV). As shown in Table IV,Group I had a significantly longer duration of stay[39∫113.8 (74.58) vs. 22∫12.4 (26.0) hours], higherAPACHE II score [9∫5.0 (9.92) vs. 3∫2.0 (3.66)],and higher TISS scores [26∫5.3 (29.23) vs.20.0∫3.8 (20.09)] than Group II.

Discussion

We have reviewed the course and outcome of 46obstetric patients admitted to our general ICUover a four-year period. The cohort we studied rep-resents 0.2% of all deliveries performed during thesame period and an ICU utilization rate of 2.3%.This low incidence of ICU admissions is also typi-cal of other series of obstetric patients (4, 5). The

Obstetric patients in General ICU 849

main admission diagnoses were obstetric con-ditions, namely excessive uterine hemorrhage andpregnancy-induced hypertension. The reason forintensive care admission was mechanical venti-lation in 41% of patients. Most of these patients(79%) underwent general anesthesia for emergencycesarean section. Mechanical ventilation was con-tinued in the postoperative period because ofhemodynamic instability in 11 patients and diffi-culty in weaning from the ventilator in a furtherfour patients. A smaller group (21%) requiredmechanical ventilation for medical indications.There can be little doubt that intensivists in an in-tensive care unit best treat these patients.

The reason for admission in the remainder ofthe patients (59%) was for monitoring and treat-ment purposes. Most of these patients (93%) werein the immediate postpartum period. While it isrecognized that many of the complications seen inthe antepartum state may improve dramaticallyafter delivery, our experience with these patientssuggests that a high level of care is often requiredin many of them. The end-points we used for de-fining optimal treatment, adequate oxygenation,urine output and blood pressure, are typical thera-peutic goals in the management of critically illICU patients. However, we believe that they applyequally to the management of this group of pa-tients. Ensuring adequate oxygenation is import-ant in all seriously ill patients. Thirty percent ofpatients in our study had significant hypoxemia onadmission, due to pneumonia, post-anestheticcomplications and pulmonary edema. All requiredhigh concentration oxygen therapy. The monitor-ing of urine output is an essential measure of ade-quate renal perfusion. This may have particularrelevance in obstetric patients with severe pre-ec-lampsia, characterized by a diastolic blood press-ure .110 mmHg, with 2π or more proteinuria andoliguria. These patients accounted for 44% of ad-missions in our series. Low cardiac filling press-ures, high systemic vascular resistance and low car-diac output characterize this condition (6). This isa consequence of altered endothelial permeability,extravasation of plasma into the interstitium andvasoconstriction. More important than simplemonitoring of urine output are the interventionsrequired to maintain adequate urine flow. In thesetting of possible underfilling of the vascular com-partment, the initial response is the giving offluids. At the same time, it is important to preventover administration of fluids as patients with se-vere pre-eclampsia have altered left ventricularcompliance (6). Excessive fluid administration maythus result in pulmonary edema. For these reasonswe monitored fluid administration by measuringand following the central venous pressure. Aggres-

C Acta Obstet Gynecol Scand 79 (2000)

sive blood pressure control is another importantend-point in the treatment of these young patientsin whom acute rises of blood pressure may posesignificant risk. Therapy is best given by continu-ous infusion of vasoactive agents, which ideally re-quires on-line measurement of intra-arterial bloodpressure to rapidly achieve end-points while avoid-ing excessive drops in blood pressure.

Measuring the central venous pressure and in-tra-arterial blood pressure are invasive proceduresand many of the obstetric complications we havedescribed may be associated with either thrombo-cytopenia (for example, the HELLP syndrome andsevere pre-eclampsia) or altered coagulation (forexample, following major blood transfusion). Thequestion may be asked whether these patients re-quire the intensity of care we provided. One wayto answer this is by examining standard severity-of-injury scores. We found that the APACHE IIscore for the total group was 6∫3.9 (7.24) and forthe monitored patients, 3∫2.02 (3.66). These fig-ures indicate a low severity of illness and predictlow mortality. APACHE scores are typically low inobstetric patients admitted to an ICU rangingfrom 6.8 to 11 in recent series (7, 8). However,many of the abnormalities commonly associatedwith complicated pregnancies, especially the preg-nancy-induced hypertension diseases, such as elev-ated liver enzymes and low platelet count, are notevaluated by the scoring system, thereby leading toa possible underestimation of severity. Neverthe-less, El-Sohl and Grant (9) found that theAPACHE II and other systems in current usecould predict severity of illness and ICU outcomein critically ill obstetric patients. Because many ofthese patients are admitted for monitoring pur-poses, it may be more relevant to assess the com-plexity of the interventions required to reach pre-determined end-points of therapy. The TISS scoremay serve this purpose. The median TISS scorefor our patients was 24∫6.1 (25.28) and for themonitored patients 20∫3.8 (20.09). This is signifi-cant as scores over 20 indicate an intensity of inter-vention requiring ICU nursing, irrespective of theseverity of illness score (2). We do not have a con-trol group of pregnant patients with similar com-plications who were not admitted to the ICU.However, the mortality rate in our series was low(2.3%). Mortality rates for obstetric patients ad-mitted to ICUs vary and may be as high as 35%(7, 10). We speculate that the low mortality rate inour series may at least be due in part to the highlevel of care available in a general ICU. It has beenshown in other settings, for example after abdomi-nal aortic aneurysm surgery, that the presence oftrained intensive care unit staff results in improvedpatient outcome (11).

850 J. Cohen et al.

Three of the patients were admitted to theICU prior to delivery (one because of an ovarianabscess, one with acute cholecystitis and onewith pneumonia) and two underwent subsequentsurgery (drainage of abscess and cholecys-tectomy). Fetal monitoring was performed rou-tinely in the ICU and obstetric decisions were al-ways made by the attending obstetrician. Twopatients underwent cesarean section due to fetaldistress and both mothers and infants made afull recovery.

In conclusion, we have shown that complicatedobstetric patients requiring either mechanical ven-tilation or invasive and intensive monitoring re-quire a high level of intervention. In view of theirlow ICU utilization rate, good prognosis and shortlength of stay, we suggest a liberal admission policyto a general ICU be encouraged for these patients.

Acknowledgments

We are grateful for the editorial and secretarial help of MrsGloria Ginzach and Mrs Melanie Kawe.

References

1. Mabie WC, Sibai BM. Treatment in an obstetric intensivecare unit. Am J Obstet Gynecol 1990; 162: 1–5.

2. Knaus WAS, Draper EA, Wagner DP, Zimmerman JE.

C Acta Obstet Gynecol Scand 79 (2000)

APACHE II: a severity of disease classification system. CritCare Med 1985; 13: 818–29.

3. Keene AR, Cullen DJ. Therapeutic intervention scoringsystem: Update 1983. Crit Care Med 1983; 11: 1–5.

4. Monaco TJ, Spielman FJ, Katz VL. Pregnant patients inthe intensive care unit. A descriptive analysis. South Med J1993; 86: 414–17.

5. Graham SG, Luxton MC. The requirement for intensivecare support for the pregnant population. Anesthesia 1989;44: 581–4.

6. Linton DM, Anthony J. Critical care management of severepre-eclampsia. Intensive Care Med 1997; 23: 248–55.

7. Lapinsky SE, Kruczynski K, Seaward GR, Farine D,Grossman RF. Critical care management of the obstetricpatient. Can J Anaesth 1997; 44: 325–9.

8. Lewinshohn G, Herman A, Leonov Y, Klinowski E. Criti-cally ill obstetrical patients: Outcome and predictability.Crit Care Med 1994; 22: 1412–14.

9. El-Sohl AA, Grant BJB. A comparison of severity of illnessscoring systems for critically ill obstetric patients. Chest1996; 110: 1299–304.

10. Kilpatrick SJ, Matthay MA. Obstetric patients requiringcritical care. A five-year review. Chest 1992; 101: 1407–12.

11. Jenckes MW, Dorman T, Garrett E, Breslow MJ, RosenfeldBA, Lipsett PA et al. Organizational characteristics of in-tensive care units related to outcomes of abdominal aorticsurgery. JAMA 1999; 281: 1310–17.

Address for correspondence:

J. Cohen, M.D.General Intensive Care UnitRabin Medical Center, Beilinson CampusPetah Tiqva 49100Israel