For discussion

Hospital Authority Board Meeting
Clinical Management and Outcome of SARS
The purpose of this paper is to update members on:- interim findings of evaluation of clinical management and clinical outcome on Severe Acute Respiratory Syndrome (SARS) international comparison of medical treatment for SARS international comparison of SARS death rates in light of various factors confounding international comparison
Background
2.
MSDC paper no. 167 on “Clinical Management of SARS- the Present and Future” was discussed at the MSDC meeting in June 2003. The paper outlined
the setting up of a central clinical information database in the Hospital Authority to
facilitate clinical management of SARS. The data of over 1700 patients has already
been captured and further data collection and analyses are still ongoing. The
ensuing paragraphs summarise the findings of the interim analysis of the available
dataset on selected topics in clinical management and outcome. Findings on a
variety of other areas including spectrum of disease, patterns of clinical presentation,
laboratory investigations and other treatment modalities will be put forward for further
discussion upon completion of the collection and analyses of the database.

Risk and Prognostic Factors of SARS

3.
Apart from the treatments received, the clinical outcome of SARS patients also depends on both individual patient factors and the severity of coronavirus infection upon presentation to medical care. Individual patient risk factors including age, sex and comorbidities were found to be significant factors among the analysed SARS patients. The interim analysis also revealed that three particular measurable clinical parameters were significantly associated with outcome. Grouped under “prognostic factors”, they included serum Lactate Dehydrogenase (LDH) level, blood neutrophil count on presentation and the lowest oxygenation level before receiving ventilatory support. The worse these parameters were, the poorer the prognosis. These prognostic factors therefore probably reflected a more severe coronavirus infection and thus less favourable outcome. 4. Data analysis for the last SARS epidemic could only be conducted with a retrospective rather than the more desirable prospective design. Identification of
this interim set of risk and prognostic factors served to partially overcome the
limitations of these retrospective studies. The effectiveness of different treatment
modalities could then be more reliably compared when the patients groups being
compared had similar characteristics in terms of the risk and prognostic factors.
Evaluation of Clinical Management

5.
Based on statistical adjustment for the risk and prognostic factors, the
use of ribavirin and the use of steroids did not appear to have an independent
favourable nor deleterious effect on patient outcome.

6.
A retrospective case-matched analysis of 44 patients showed that early treatment with a protease inhibitor combination drug called Kaletra, which has been
used for HIV infection, in addition to ribavirin/steroid, was associated with a
significantly lower death rate, lower intubation rate, and lower steroid dosage.
A randomized placebo-control study for the use of Kaletra and ribavirin in SARS is
called for in future epidemics.
7.
Another retrospective case-matched analysis of nearly 300 patients showed that there was no difference in death rates for patients who received no
steroids versus those who received moderate to high dose steroids. Patients
who did not receive steroids had longer hospital stay before discharge or death.
Despite a lack of difference in eventual outcome, those who received higher doses of
steroids were sicker patients at the time of treatment initiation. This suggests a
beneficial effect of higher doses of steroids for the sicker patients, and a randomized
control study could be considered during the next epidemic.
Evaluation of Clinical Outcome

8.
Hospital-specific death rates were difficult to ascertain as there were nearly 200 patients who were hospitalized in two or more hospitals during their
SARS illness. Applying a hospital assignment method to these patients, the crude
death rates of individual hospitals varied from 0% to 79%. However, after taking
into consideration the specific case mix of each hospital population (i.e. the
composition of the population with regard to the risk and prognostic factors), there
was no statistical significance in the adjusted death rates across all HA
hospitals.
International Comparison of Medical Treatment of SARS
9.
A literature review on the treatment protocols used in different regions shows in addition to supportive therapy, broad-spectrum antibiotics were used invariably in all regions. Ribarivin was used in 88% of Canadian patients, 7% of
Singapore patients and 25% in Mainland China. Corticosteroids were used in 40%
of Canadian patients. In Singapore, 70% of ICU patients received steroids and 4%
of non-ICU patients did. In Mainland China, more than half of the patients were
given either maintenance or pulse steroid therapy. As there were wide variations in
the clinical indications for the use of steroids, as well as in the dosages used and
duration of usage, treatment outcome from the use of steroids cannot be compared
among regions. All in all, no conclusion can be drawn as to a standard
treatment for SARS based on international experience.
International Comparison of SARS Death Rates
10.
According to WHO figures up to 7 August 2003, the case fatality rate (CFR) for Hong Kong was 17%, in comparison to 17% for Canada, 7% for China,
27% for Taiwan, 14% for Singapore, and 8% for Vietnam. The variation in CFRs in
different regions of the world may appear alarming, but must be viewed with
caution as there are quite a number of confounding variables which may
render such inter-regional comparison not interpretable. The confounding
variables include the case mix (i.e. the composition of the patient population with
regard to age, sex, chronic health status, severity of coronavirus infection, etc.), as
well as dilutional factors which may influence the denominator, i.e. the total SARS
cohort, in calculating the death rates.
11.
The age distribution among SARS patients differed among
regions. Hong Kong and Taiwan shared similar age profile in that there was a larger
proportion of elderly patients (e.g. 15-20% above age 70) as compared to other
regions such as China and Singapore where there was a predominance of younger
patients (e.g. 5-6% above age 70).
12.
The case mix of different SARS populations also differed greatly
in the proportion of infected healthcare workers (HCWs). The proportion of
infected HA HCWs was 19% for Hong Kong, as compared to 43% for Canada, 19%
for China, 41% for Singapore, and 57% for Vietnam according to WHO figures up to
7 August 2003. HCWs tended to have a survival advantage as there were more
female, younger and healthier persons. They might also present with milder disease.
The higher proportion of HCWs in the cohort would favorably lower the CFR.
13.
The lower CFRs in China in comparison to Hong Kong data might
also be partly attributed to the use of non-invasive ventilation (NIV) in critically
ill SARS patients. NIV was used in only 2% of SARS patients in Hong Kong,
compared to 13.8-32.6% in Chinese published series. The use of Chinese Medicine
might also be another factor. Further confirmation of this proposition awaits detailed
scientific analysis.
14.
Dilutional factors also contributed to an apparent difference in
CFR. Since SARS was a clinical diagnosis without need for virological confirmation,
the behaviour of SARS reporting might vary from region to region depending on
prevailing practice. An example in point was that only 12% of those initially reported in US were subsequently confirmed virologically to be SARS patients, suggesting a possibly loose clinical diagnosis and broad denominator. In Hong Kong, virologically confirmed SARS took up about 60% of our cohort, suggesting a more vigorous clinical diagnosis and thus a more stringent denominator. 15. The timing of reporting of death rates also confounded inter-
regional comparison. Examining the ICU patients reported in the literature,
mortality figures for HA ICU patients compared favourably with the Canadian and
Singaporean cohorts. The 28-day mortality after symptom onset was 22.3% for
Hong Kong as compared to the published data showing 37% for Singapore. The 28-
day mortality after ICU admission was 26.7% for Hong Kong as compared to
published data showing 34% for Toronto. The 13-week mortality was 43% versus
52.2% for Singapore.

Advice sought
16.
Members are requested to note the findings on clinical management and outcome of SARS and the international comparisons highlighted.

Hospital Authority
HAB\PAPER\66
22 September 2003

Source: http://www.hospitalauthority.hk/haho/ho/cad_bnc/123698e.pdf