Identifying predictive factors for admitting patients with severe pre-eclampsia to intensive care unit

BACKGROUND

Traditional obstetric units are inadequate for the level of monitoring required in pre-eclamptic patients. It remains to be determined which facility and tools should be implemented. The aim of this work was to identify predictive factors of admission to Intensive Care Unit (ICU) admissions for pregnancy-related hypertensive complications.

METHODS

We conducted an observational, retrospective multicenter study (Toulouse, Nantes). Both have a level III maternity unit and an ICU. The selected patients had one or more of the following diagnoses on admission in the ICU or during hospitalization: pre-eclampsia, eclampsia, HELLP syndrome, Acute Fatty Liver of Pregnancy (AFLP), Hemolytic Uremic Syndrome (HUS). SAPS II, SOFA and APACHE II on admission, and a validated nursing workload assessment score: TISS 28, were collected.

RESULTS

211 parturient women were included. According to the multivariate analysis: APACHE II and SAPS 2 severity scores >15 were significantly higher in the TISS 28 ≥ 20 group. There were also higher rates of uricemia >360 mmol/l.

CONCLUSIONS

To date, there are no reliable and validated predictive factors of severity to guide the transfer of pre-eclamptic patients to an ICU. The combination of an increased APACHE II score and uricemia, as well as a high care workload score that could help with the transfer of high-risk pre-eclamptic patients to a specific care facility. This hypothesis should be tested prospectively. This work could incite reflection on the value of creating obstetric intensive care units, according to the size of the maternity unit.

Benefits of smart pumps for automated changeovers of vasoactive drug infusion pumps: a quasi-experimental study

BACKGROUND

Manual changeover of vasoactive drug infusion pumps (CVIP) frequently lead to haemodynamic instability. Some of the newest smart pumps allow automated CVIP. The aim of this study was to compare automated CVIP with manual 'Quick Change' relays.

METHODS

We performed a prospective, quasi-experimental study, in a university-affiliated intensive care unit (ICU). All adult patients receiving continuous i.v. infusion of vasoactive drugs were included. CVIP were successively performed manually (Phase 1) and automatically (Phase 2) during two 6-month periods. The primary endpoint was the frequency of haemodynamic incidents related to the relays, which were defined as variations of mean arterial pressure >15 mm Hg or heart rate >15 bpm. The secondary endpoints were the nursing time dedicated to relays and the number of interruptions in care because of CVIP. A multivariate mixed effects logistic regression was fitted for analytic analysis.

RESULTS

We studied 1329 relays (Phase 1: 681, Phase 2: 648) from 133 patients (Phase 1: 63, Phase 2: 70). Incidents related to CVIP decreased from 137 (20%) in Phase 1 to 73 (11%) in Phase 2 (P<0.001). Automated relays were independently associated with a 49% risk reduction of CVIP-induced incidents (adjusted OR=0.51, 95% confidence interval 0.34-0.77, P=0.001). Time dedicated to the relays and the number of interruptions in care to manage CVIP were also significantly reduced with automated relays vs manual relays (P=0.001).

CONCLUSIONS

These results demonstrate the benefits of automated CVIP using smart pumps in limiting the frequency of haemodynamic incidents related to relays and in reducing the nursing workload.