Hasan Surya, Sudirgo
(2010)
Non-invasive ventilation in adult population in intensive care unit Hospital Universiti Sains Malaysia:
a retrospective study.
Masters thesis, Universiti Sains Malaysia.
Abstract
The aims of this study were to review the practice of application ofNIV
in adult patients admitted in general ICU HUSM and to investigate factors that
contributed to its success and failure. A retrospective study was done to review the NIV practice in HUSM.
We defined the characteristics of adult patients 18 years old and over who were
admitted in ICU in the period of 1 January 2008 to 31 December 2008 (1 year), who
had been put on NIV in any course of their staying in ICU. Data were collected from
patient's medical record, including the patient's ICU chart. Parameters taken were
demographic, diagnosis, patients' co-morbidity, SAPS II point, length of ICU stay,
and patients' condition when discharged from ICU: survived or not survived. The
SAPS II point was generated from the first 24 hour of patient's ICU admission. At the
time when the NIV were being applied, specific data such as arterial blood gas
analysis (ABG) values, status of organ failure and kind of supports (inotropes and
vasopressors, haemodialysis etc.) that patients received were recorded. Subsequent
evaluation on NIV setting and the ABG were looked at certain interval: at 15
\ 6
1
h, 12
1
h
and 241h hour of NIV application. If there was no ABG taken during those specific
hours, the previous or next ABG within 2 hours duration was taken, or if there was
none, was recorded as missing data. The duration of NIV in hours and the final
outcome of each patient were documented. Any deleterious events happened during
the course of NIV and the action taken by the attending doctors were also recorded. Deeper look on the attending doctors' note was taken for the patients who had failed
NIV and needed intubation or reintubation. Descriptive analysis was used to describe
the patients' characteristics. Relevant data were analysed using non-parametric test:
Kruskal-Wallis test, Chi-Square test or Mann-Whitney test, which ever appropriate. We categorized 54 patients into either de novo respiratory failure (de novo),
acute on chronic respiratory failure (AOC) or acute cardiogenic pulmonary oedema
(ACPO). The proportion of de novo was high (74.1 %) when the other two were low
( 14.8% for AOC and 11.1% for ACPO). Among those three groups, there were no
significant differences in survival rates (77.5%, 87.5% and 83.3% in de novo, AOC
and ACPO respectively). ICU stay was significantly longer in de novo groups
(11.5±8.04 days vs. 4.9±1.46 days in AOC and 6.8±5.19 days in ACPO; p-value =
0.029). When we looked further at patients in de novo group who were diagnosed as
pneumonia, the mortality and the NIV failure rate were significantly higher than the
non-pneumonia patients (50% vs 10.7% mortality rate; p-value = 0.004, and 66.7% vs.
39.3% NIV failure rate; p-value = 0.017). In addition, when patients failed NIV and
need intubation/re-intubation, the ICU stay would be doubled (15.1±8.5 days vs.
7.5±5.83 days). The overall mortality rate didn't deviate from the predicted mortality
rate calculated from the SAPS II point, and it was clearly shown that there was a
significant difference in SAPS II point between the survivors and non survivors
(SAPS II point 35.7±13.39 vs. 46.1±15.63 ; p-value = 0.036). However, there is no
correlation between SAPS II point and NIV failure rate. In our ICU, most patients treated with NIV were de novo
respiratory failure. There were no significant differences in mortality and NIV failure
rates among different groups of respiratory failure, although ICU stay was clearly longer in de novo group. ICU stay would also be longer in those who failed NIV and
need intubation/re-intubation. SAPS II point accurately predicted mortality rate but
not NIV failure rate. Pneumonia patients in de novo group had significantly higher
mortality rate as well as NIV failure rate compared to non-pneumonia patients. Indeed,
among all of the analysed variables, pneumonia was the only one that had strong
association with NIV failure. We concluded that our NIV practice results were not
consistent with some reports in other clinical trials, except that pneumonia was a
contributing factor for NIV failure and that those who failed NIV would have longer
ICU stay. But we also emphasized that differences might exist between controlled
clinical trials and real clinical practice as this retrospective study basically was. From
this retrospective study alone, it was not possible to determine whether our NIV
practice had complied with the best standard practices. The NIV utilization was low
and it was expected that had the medical officers (MOs) been more confident, more
patients should have been treated with NIV. Implementation of "Immediate NIV" and
"alternate NIV" were considered as good practices, but the high proportion of de novo,
including pneumonia patients treated with NIV suggested that the MOs might not be
aware about the results from the previous studies which encouraged careful selection
for that kind of patients. The presence of patients who died while on NIV would also
speculate that intubation should have been taken earlier. The absence of NIV
guidelines in our institution might leave improper decisions to be widely opened, and
the status as a teaching hospital might become a disadvantage as new MOs with less
experience would come and those experienced ones would go. Therefore, we believed
that the availability of standard protocol and proper training were essential for future
improvement in NIV application in our institution.
Actions (login required)
 |
View Item |
