Objective Palliative Prognostic Score among Advanced Cancer Patients
Yen-Ting Chen1,3, Chih-Te Ho1,3,4, Hua-Shai Hsu1,3,4, Po-Tsung Huang1,3, Chin-Yu
Lin2, Chiu-Shong Liu1,4,5, Tsai-Chung Li6,7, Cheng-Chieh Lin1,4,5,7,Wen-Yuan Lin1,3,4,5
1
Department of Family Medicine, 2Department of Nursing, and 3Hospice Palliative
Medicine Unit, China Medical University Hospital, Taichung, Taiwan
4
School of Medicine, 5Graduate Institute of Clinical Medical Science and 6Graduate
Institute of Biostatistics, China Medical University, Taichung, Taiwan;
7
Institute of Health Care Administration, College of Health Science, Asia University,
Taichung, Taiwan;
Co-Author e-mail addresses:
Y.T. Chen: jamesytchen@gmail.com
C.T. Ho: ctho@mail.cmuh.org.tw
H.S. Hsu: lavendertea@pchome.com.tw
P.T. Huang: d16973@mail.cmuh.org.tw
C.Y. Lin: n4795@mail.cmuh.org.tw
C.S. Liu: liucs@ms14.hinet.net
T.C. Li: tcli@mail.cmu.edu.tw
C.C. Lin: cclin@mail.cmuh.org.tw
Correspondence and reprint request to:
Wen-Yuan Lin M.D., M.S., Ph.D.
Department of Family Medicine, China Medical University Hospital,
2, Yuh-Der Road, Taichung, Taiwan 404
1
Tel: +886-4-22052121 ext 4507, Fax: +886-4-22361803,
E-mail: wylin@mail.cmu.edu.tw
Number of tables : 3
Number of figures : 2
Number of references : 32
Word count : 2636 (references no included)
2
Abstract
Context. The accurate prediction of survival is one of the key factors in the
decision-making process for patients with advanced illnesses.
Objectives. This study aims to develop a short-term prognostic prediction method
involving such objective factors, past medical history, vital signs, and blood tests, for
use with advanced cancer patients.
Methods. Medical records gathered upon admission of advanced cancer patients in
the Palliative Care Unit at a tertiary hospital in Taiwan were retrospectively reviewed.
The records included
demographics, history of cancer treatments, performance
status, vital signs and biological parameters, Multivariate Cox regression analyses and
receiver operating characteristic (ROC) curve employed for model development.
Results. The Objective Palliative Prognostic Score (OPPS) was determined by using
six objective predictors identified by multivariate Cox regression analysis. The
predictors include heart rate > 120/min, WBCs > 11000/mm3, platelets < 130000/mm3,
serum creatinine level > 1.3 mg/dl, serum potassium level > 5 mg/dl, and no history
of chemotherapy. The area under the ROC curve used to predict 7 days survival was
82.0% (95% CI: 75.2 to 88.8%). If any three predictors out of the six were reached,
survival within 7 days was predicted with 68.8% sensitivity, 86.0% specificity, 55.9%
positive predictive value (PPV), and 91.4% negative predictive value (NPV).
Conclusions. The OPPS consisted of six objective predictors for the estimation of 7
days survival among advanced cancer patients and showed a relatively high accuracy,
specificity, and NPV. Objective signs, such as vital signs and blood tests, may help
3
clinicians make decisions at the very late stages of the life of the patients.
4
Keywords
advanced cancer, palliative care, prognosis, survival, prediction
5
Running Title
objective palliative prognostic score
6
Introduction
Decision making is an important issue in palliative care, especially as patients approach the
end of their lives. For patients, families, and physicians, timing the shift from life-prolonging
approaches to more palliative ones that focus on quality of life and comfort, is challenging.
The accurate prediction of survival is one of the key factors in the decision-making process
from goal setting in the early stages of palliative care, which include care giving or
expectations of care, to ethical issues at the end of a patient’s life; such ethical issues include
withholding and withdrawing treatment, palliative sedation, and nutrition support.1 However,
traditional prediction by physicians is unreliable, and most physician prognoses tend to be
both imprecise and overly optimistic.2, 3
Many previous studies have formulated survival predictions according to a variety of items,
including performance status, symptoms, and biological parameters.4-9 These predictions
could significantly help physicians in decision making and in providing patients and their
families with useful information for goal setting.6, 10, 11 However, most of these predictions
include subjective factors, such as patient symptoms and performance status,12-16 which may
be influenced by treatment, time, and physicians. These subjective factors may also be
differently evaluated by each caregiver and may be estimated with difficulty by junior
physicians. Thus, the idea of being able to create a prognosis from objective factors, such as
vital signs and blood test, is appealing to clinicians.
A previous study conducted in Italy revealed that white blood cell (WBC), lymphocyte
percentage, and pseudocholinesterase levels were independent objective predictors of
7
survival.17 Other studies have also been conducted recently. These studies found that some
objective factors, such as leukocytosis, lymphocytopenia, thrombocytopenia, elevated
C-reactive protein (CRP), decreased albumin level, high lactate dehydrogenase level, high
blood urea nitrogen (BUN) level, and high respiratory rate, have prognostic significance in
shorter survival.18-23
Decision making in the very late stages of life has more ethical concerns and is more
patient-centered, which makes it more difficult for physicians. The diagnosis of death is
central to optimal decision making at the end of patient life. However, diagnosing death is
often a complex process that has many obstacles, which lend difficulty to the clear
recognition of the dying phase.24 Thus, increasing interest has been directed toward the
prediction of short-term survival.14, 19, 21, 25, 26 Chiang et al. constructed a 7 days survival
prediction model based on cognitive status, edema, Eastern Cooperative Oncology Group
(ECOG) performance status, BUN and respiratory rate with a good area under the curve
(AUC) of 0.81 (P<0.001, 95% CI: 0.76 to 0.86).19 Ohde et al. also suggested a two-week
prognostic prediction model for terminal cancer patients made up of 5 items: anorexia,
dyspnea, edema, BUN > 25.0 mg/dl, and platelet < 260,000/mm3, with a good AUC of 0.83
(0.75-0.91).21 However, both studies still have subjective objects that may potentially be
judged by physicians inaccurately.
This study primarily aims to develop a validated short-term prognostic prediction
consisting objective factors to facilitate the decision making of advanced cancer patients, their
relatives, and physicians in the very late stages of life.
8
Methods
In this cross-sectional study, cases advanced cancer patients admitted to the Hospice
Palliative Care Unit in a tertiary hospital in Taiwan from June 2005 to September 2007 were
retrospectively analyzed. The inclusion criteria included patients aged 20 years and above
who were admitted to the Hospice Palliative Care Unit of the China Medical University
Hospital, have a complete medical record for vital signs, and routinely had basic blood test
upon admission. For the patients who were admitted more than twice, the first admission was
selected for use in the study. A total of 240 patients were enrolled in this study, and six were
excluded because of the lack of survival data. Finally, 234 advanced cancer patients were
selected. The study was approved by the ethics committee of the China Medical University
Hospital.
All potential objective prognostic predictors, including demographic data, diagnostic and
therapeutic information, health performance status, vital signs, and biological parameters,
were collected by experienced hospice physicians and registered nurses within 24 hours of
hospice admission. The potential parameters were based on previous studies or on the clinical
experience of physicians. Demographic data include gender, age, and body mass index (BMI).
Diagnostic and therapeutic information include the primary site of malignancy and metastasis
and history of radiation therapy or chemotherapy. Health performance status was evaluated
and recorded through the ECOG performance status. Vital signs recorded at admission
include blood pressure, heart rate, and oxygen usage. Laboratory tests were conducted
through routine blood drawing upon admission. The evaluated parameters include WBC,
9
hemoglobin (Hb), platelets, alanine aminotransferase (ALT), aspartate aminotransferase
(AST), BUN, creatinine (Cr), total bilirubin, albumin, CRP, potassium (K), sodium (Na) and
calcium (Ca). Survival days were defined as the period from the date of admission to the
Hospice Palliative Care Unit to the date of death. Patients who survived for more than 7 days
were labeled as “long survivors,” and those who died within 7 days were labeled as “short
survivors.”
Statistical analysis
Descriptive statistics were summarized as frequencies and percentages for categorical
variables, and continuous variables were expressed as mean and standard deviation (SD).
Vital signs and biological parameters were divided into groups based on average, the normal
range of our laboratory, the clinical experience of the physicians, and previous studies. 18, 19,
21-23, 27
Student’s t test and Chi-square test were used to reveal differences between long survivors
and short survivors. All the candidates for objective predictors with p-value < 0.1 in
univariate analyses were selected for the multivariate Cox regression analyses. The
independent predictors with p-value < 0.05 in the multivariate analysis comprised final
prognostic model, and each score was based on the odds ratio of every predictor. The AUC
and survival rate at each point were obtained from the ROC curve. All statistical tests were
two sided, with the significance level at 0.05. These statistical analyses were performed using
the MAC version of the SPSS statistical software (20th version, SPSS Inc., Chicago, IL,
USA)
10
Results
Table 1 shows the baseline characteristics between the long survival and short survival
groups. Of the 234 patients, 48 (20.5%) subjects were categorized as short survivors (survival
within 7 days), whereas 186 (79.5%) subjects were long survivors. Both groups had slightly
more males than females, and the majority of the primary site of malignancy is
gastrointestinal. No significant differences were observed in terms of age, BMI, or ECOG
between groups. The variables that were more discriminant for shorter survival in the
univariate analysis (P<0.05) were brain metastasis and no history of chemotherapy.
Table 2 presents the results of the univariate analysis of vital signs and biological
parameters. Heart rate > 120 bpm, use of oxygen, WBC > 11000/mm3, platelets <
130000/mm3, ALT > 40 mg/dl, BUN > 26 mg/dl, Cr > 1.3 mg/dl, bilirubin > 2 md/dl, CRP > 8
mg/dl and K > 5.0 mg/dl are related to short-survival. Hb, albumin, Na, and Ca were not
associated with the length of survival.
Table 3 demonstrates the results of backward stepwise Cox regression analyses. The
determinants of patients who died within 7 days (short survival) are no history of
chemotherapy, HR >120 bpm, WBC > 11000/mm3, platelet < 130000/mm3, Cr > 1.3mg/dl,
and K > 5 mg/dl. The determinants were subsequently selected as prognostic predictors based
on a p-value <0.05. The final models of prognostic prediction were employed to establish the
Objective Palliative Prognostic Score (OPPS) and ROC curves (Figure 1). The AUC was
82.0% (95% CI: 75.2 to 88.8%). When total scores are 0 to 2 points, 3 points, or 4 to 5 points,
one-week mortality rates are 8.6%, 50% and 73.3%, respectively (Figure 2). If any three
11
predictors out of the six were reached, survival within 7 days was predicted a sensitivity of
68.8%, specificity of 86.0%, positive predictive value of 55.9% and negative predictive value
(NPV) of 91.4%.
12
Discussion
In this study, we proposed a prognostic prediction model, the OPPS, for estimating the
probability of 7 days survival among advanced cancer patients. To our knowledge, OPPS is
the first prognostic survival prediction model exclusively composed of objective factors. The
six predictors identified by the multivariate Cox regression analyses are (1) no history of
chemotherapy, (2) heart rate > 120 bpm, (3) WBC > 11000/mm3, (4) platelets < 130000/mm3,
(5) Cr > 1.3 mg/dl, and (5) K > 5.0 mg/dl. In our study, the AUC showed relatively good
accuracy (AUC: 0.82, p < 0.001, 95% CI: 0.75 to 0.89). The OPPS also showed high
specificity (86%) and high NPV (91.4%) for predicting 7 days survival when 3 predictors
were reached.
Hemodynamic statuses are routinely recorded routinely as important clinical signs, and
hemodynamic changes, such as drop of blood pressure, tachycardia or bradycardia, are
frequently noted during the dying process. Tachycardia was a significantly independent
predictor noted in this study and was also reported by Elsayem and Mercadante et al. for
short-term survival among terminal cancer patients.20, 23 Lower blood pressure was not a
significant factor in our analysis, though Mercadante et al. reported lower systolic blood
pressure as a significant factor for shorter survival in advanced cancer patients who received
home care service.23 By contrast, Elsayem et al. reported that diastolic blood pressure was not
a significant factor, similar to our model, which warrants further investigation.20
Leukocytosis, thrombocytopenia, and elevated serum Cr and K levels, as biological
parameters, are independent prognostic factors in this multivariate analysis and are included
13
in the formula for OPPS. Leukocytosis is a well-known survival predictor in advanced cancer
patients.4, 6, 10 In our analysis, leukocytosis also worked well in short-term survival prediction.
Thrombocytopenia is common in terminal patients because of the impaired production of
platelets caused by the decrease or absence of megakaryocytes owing to the myelosuppressive
effects of chemotherapy or radiation therapy or as a result of the infiltration of the bone
marrow by malignant cells.28 Ohde et al. also suggested that the members of the low platelet
group of terminal cancer patients are more likely to die within two weeks.21 Impaired renal
function is another common condition among terminal cancer patients. Both Chiang and Ohde
et al. included elevated serum BUN level in their models as a predictor for short-term
survival.19 In our multivariate analyses, elevation in BUN and Cr are both independent
predictors for patients that died within 7 days. Including elevated Cr level in the formula
results in better AUC than including elevated BUN level. This study is unique because of the
inclusion of hyperkalemia in the OPPS. Electrolyte imbalance, a very common condition in
advanced cancer patients, has never been included in previous prognostic models.
Hyperkalemia was found to be an accurate predictor for 7 days survival.
One notable finding in our study is that no history of chemotherapy is also included in our
prognostic model. Patients who have received chemotherapy might have better survival.29-31
However, no study has proven that chemotherapy could be used in the prediction of
short-term survival. This topic warrants further investigation.
The accuracy prediction of survival is very important in palliative care sot that appropriate
treatment can be arranged and discussion with the patients and their families about goal
14
setting can be facilitated. Clinical prognoses are mostly based on physician experience and the
observation of the clinical condition of the patient. However, the causes of death in advanced
cancer patients vary, and not every patient has a smooth downward trajectory. Although
terminal symptoms, such as dyspnea, anorexia, and poor performance, are very strong
predictors for the dying process,7,
9, 32
terminal patients who are near death may exhibit
minimal clinical change in very late stages of life. Clinicians often find great difficulty in
making treatment decisions in the very late stages of life of the patients, especially in
performing invasive procedures or withdrawing life-sustaining therapy. A useful prognostic
tool for the prediction of short-term survival may help guide treatment. Chiang et al.
established a prognostic 7 days survival formula for patients with terminal cancer patients.
This formula consisted three subjective clinical findings and two objective records (BUN and
respiratory rate). The formula also had good accuracy in AUC (0.81, p < 0.001, 95% CI: 0.76
to 0.86) and high NPV (91.7%), which are similar to the findings of this study. However, the
predictor model established based on a complicated formula can not be easily applied to the
clinical setting.
This study has some limitation. The final prognostic prediction model retrospectively
derived from a single institution with a limited numbers of patients may have a limited nature
and may be over-fitted, and the degree and criteria of patients admitted to palliative care differ
around the world. Patients are admitted to our Hospice Palliative Care Unit mostly because of
acute problems, such as pain, infection, electrolyte imbalance, or conscious disturbance. Thus,
blood sampling upon admission is usually performed. However, blood sampling, being an
15
invasive procedure that is uncomfortable for patients, may not be routinely performed in some
palliative care facilities. Moreover, this prognostic score is solely applied to advanced cancer
patients and may not be applicable to non-cancer patients. These findings require further
evaluation in larger prospective studies.
In conclusion, we suggest a 7 days prognostic prediction score for advanced cancer patients.
This score consists of six objective factors, namely, heart rate > 120 bpm, WBCs >
11000/mm3, platelets < 130000/mm3, Cr > 1.3 mg/dl, K > 5 mg/dl, and no history of
chemotherapy. The OPPS has a good AUC (82.0%, 95% CI: 75.2 to 88.8%). When three
predictors were reached, death within 7 days was predicted with relatively good specificity
(86.0%) and NPV (91.4%). OPPS is the first prediction model for survival of the advanced
cancer patients that is made up of only objective subjects. The findings of this study may
provide useful information for the development of new prognostic models.
16
Disclosures and Acknowledgements
This study was financially supported by grants from the National Science Council of
Taiwan (NSC94-2314-B-039-025, NSC95-2314-B-039-008, and NSC96-2314-B-039-015)
and China Medical
University Hospital (DMR-99-109, and
DMR-101-058, and
DMR-102-065).
We thank the medical staff and all participants in hospice palliative medicine for their
assistance in completing this study.
17
References
1. Weissman DE. Decision making at a time of crisis near the end of life. JAMA
2004;292:1738-43.
2. Glare P, Virik K, Jones M, et al. A systematic review of physicians' survival predictions
in terminally ill cancer patients. BMJ 2003;327:195-8.
3. Gripp S, Moeller S, Bölke E, et al. Survival prediction in terminally ill cancer patients by
clinical estimates, laboratory tests, and self-rated anxiety and depression. Journal of
clinical oncology : official journal of the American Society of Clinical Oncology
2007;25:3313-20.
4. Viganó a, Bruera E, Jhangri GS, et al. Clinical survival predictors in patients with
advanced cancer. Archives of internal medicine 2000;160:861-8.
5. Viganò a, Dorgan M, Buckingham J, Bruera E, Suarez-Almazor ME. Survival prediction
in terminal cancer patients: a systematic review of the medical literature. Palliative
medicine 2000;14:363-74.
6. Maltoni M, Caraceni A, Brunelli C, et al. Prognostic factors in advanced cancer patients:
evidence-based clinical recommendations--a study by the Steering Committee of the
European Association for Palliative Care. Journal of clinical oncology : official journal of
the American Society of Clinical Oncology 2005;23:6240-8.
7. Stone PC, Lund S. Predicting prognosis in patients with advanced cancer. Annals of
oncology : official journal of the European Society for Medical Oncology / ESMO
2007;18:971-6.
18
8. Glare P, Sinclair C, Downing M, et al. Predicting survival in patients with advanced
disease. European journal of cancer 2008;44:1146-56.
9. Ripamonti CI, Farina G, Garassino MC. Predictive models in palliative care. Cancer
2009;115:3128-34.
10. Maltoni M, Nanni O, Pirovano M, et al. Successful validation of the palliative
prognostic score in terminally ill cancer patients. Italian Multicenter Study Group on
Palliative Care. Journal of pain and symptom management 1999;17:240-7.
11. Morita T, Tsunoda J, Inoue S, Chihara S. Improved accuracy of physicians' survival
prediction for terminally ill cancer patients using the Palliative Prognostic Index.
Palliative Medicine 2001;15:419-24.
12. Pirovano M, Maltoni M, Nanni O, et al. A new palliative prognostic score: a first step
for the staging of terminally ill cancer patients. Italian Multicenter and Study Group on
Palliative Care. Journal of pain and symptom management 1999;17:231-9.
13. Anderson F, Downing GM, Hill J, Casorso L, Lerch N. Palliative performance scale
(PPS): a new tool. Journal of Palliative Care 1996;12:5-11.
14. Caraceni A, Nanni O, Maltoni M, et al. Impact of delirium on the short term prognosis
of advanced cancer patients. Italian Multicenter Study Group on Palliative Care. Cancer
2000;89:1145-9.
15. Chuang R-B, Hu W-Y, Chiu T-Y, Chen C-Y. Prediction of survival in terminal cancer
patients in Taiwan: constructing a prognostic scale. Journal of pain and symptom
management 2004;28:115-22.
19
16. Yun YH, Heo DS, Heo BY, et al. Development of terminal cancer prognostic score as an
index in terminally ill cancer patients. Oncology Reports 2001;8:795-800.
17. Maltoni M, Pirovano M, Nanni O, et al. Biological indices predictive of survival in 519
Italian terminally ill cancer patients. Italian Multicenter Study Group on Palliative Care.
Journal of Pain and Symptom Management 1997;13:1-9.
18. Kikuchi N, Ohmori K, Kuriyama S, et al. Survival prediction of patients with advanced
cancer: the predictive accuracy of the model based on biological markers. Journal of pain
and symptom management 2007;34:600-6.
19. Chiang J-K, Lai N-S, Wang M-H, Chen S-C, Kao Y-H. A proposed prognostic 7-day
survival formula for patients with terminal cancer. BMC public health 2009;9:365.
20. Elsayem A, Mori M, Parsons Ha, et al. Predictors of inpatient mortality in an acute
palliative care unit at a comprehensive cancer center. Supportive care in cancer : official
journal of the Multinational Association of Supportive Care in Cancer 2010;18:67-76.
21. Ohde S, Hayashi A, Takahasi O, et al. A 2-week prognostic prediction model for
terminal cancer patients in a palliative care unit at a Japanese general hospital. Palliative
medicine 2011;25:170-6.
22. Partridge M, Fallon M, Bray C, et al. Prognostication in advanced cancer: a study
examining an inflammation-based score. Journal of pain and symptom management
2012;44:161-7.
23. Mercadante S, Valle A, Porzio G, et al. Prognostic factors of survival in patients with
advanced cancer admitted to home care. Journal of pain and symptom management
20
2013;45:56-62.
24. Solomon MZ, O'Donnell L, Jennings B, et al. Decisions near the end of life:
professional views on life-sustaining treatments. American journal of public health
1993;83:14-23.
25. Bozcuk H, Koyuncu E, Yildiz M, et al. A simple and accurate prediction model to
estimate the intrahospital mortality risk of hospitalised cancer patients. International
journal of clinical practice 2004;58:1014-9.
26. Stone P, Kelly L, Head R, White S. Development and validation of a prognostic scale
for use in patients with advanced cancer. Palliative Medicine 2008;22:711-7.
27. Suh SY, Ahn HY. Lactate dehydrogenase as a prognostic factor for survival time of
terminally ill cancer patients: a preliminary study. European Journal of Cancer
2007;43:1051-9.
28. Avvisati G, Tirindelli MC, Annibali O. Thrombocytopenia and hemorrhagic risk in
cancer patients. Critical Reviews in Oncology/Hematology 2003;48:S13-S16.
29. Gebski V, Burmeister B, Smithers BM, et al. Survival benefits from neoadjuvant
chemoradiotherapy or chemotherapy in oesophageal carcinoma: a meta-analysis. The
lancet oncology 2007;8:226-34.
30. Scher HI, Fizazi K, Saad F, et al. Increased survival with enzalutamide in prostate
cancer after chemotherapy. The New England journal of medicine 2012;367:1187-97.
31. Sjoquist KM, Burmeister BH, Smithers BM, et al. Survival after neoadjuvant
chemotherapy or chemoradiotherapy for resectable oesophageal carcinoma: an updated
21
meta-analysis. The lancet oncology 2011;12:681-92.
32. Innes S, Payne S. Advanced cancer patients' prognostic information preferences: a
review. Palliative medicine 2009;23:29-39.
22
Table 1: Baseline characteristics between short and long survivors
Short-survival
Long-survival
Patient number (%)
48 (20.5)
186 (79.5)
Male, n (%)
31 (64.6)
100 (53.8)
0.195
Age, mean (SD)
62.8 (13.6)
62.9 (14.2)
0.783
BMI, mean (SD)
19.6 (3.4)
19.7 (3.5)
0.909
ECOG, n (%)
p-value
0.703
1
2 (4.2)
11 (5.9)
2
13 (27.1)
49 (26.3)
3
13 (27.1)
63 (33.9)
4
20 (41.7)
63 (33.9)
Primary site, n (%)
0.660
GI
21 (43.8)
75 (40.3)
Lung
5 (10.4)
37 (19.9)
GU
2 (4.2)
11 (5.9)
Head and neck
8 (16.7)
28 (15.1)
Breast
2 (4.2)
8 (4.3)
others
10 (20.8)
27 (14.5)
Liver
14 (29.2)
58 (31.2)
0.862
Lung
15 (31.2)
59 (31.7)
1.000
Metastasis, n (%)
23
Bone
17 (35.4)
69 (37.1)
0.868
Brain
3 (6.2)
36 (19.4)
0.030
Lymph
7 (14.6)
35 (18.8)
0.673
Abdomen
4 (8.3)
10 (5.4)
0.494
History of chemotherapy, n (%)
21 (43.8)
117 (62.9)
0.021
History of Radiotherapy, n (%)
22 (45.8)
94 (50.5)
0.628
Abbreviations: BMI, body mass index; ECOG, Eastern Cooperative Oncology Group
performance status; GI, gastrointestinal; GU, genitourinary.
24
Table 2: Vital signs and biological parameter data between short and long survivors.
Short-survival
Long-survival
p-value
SBP < 100 mmHg, n (%)
8 (16.7)
26 (14.0)
0.648
HR > 120 BPM, n (%)
19 (39.6)
33 (17.7)
0.003
Use of oxygen, n (%)
29 (60.4)
79 (42.7)
0.034
Ascites, n (%)
8 (16.7)
22 (11.8)
0.345
WBC > 11000/mm3, n (%)
35 (72.9)
75 (40.3)
< 0.001
Hb < 10 g/dl, n (%)
19 (39.6)
81 (43.5)
0.744
Platelet < 130000/mm3, n (%)
17 (35.4)
33 (17.7)
0.011
ALT > 40 U/L, n (%)
10 (20.8)
13 (7.0)
0.011
BUN > 26 mg/dl, n (%)
26 (54.2)
46 (24.7)
< 0.001
Cr > 1.3 mg/dl, n (%)
25 (52.1)
37 (19.9)
< 0.001
Total bilirubin > 2 mg/dl, n (%)
16 (33.3)
34 (18.3)
0.030
Albumin < 2 g/dl, n (%)
23 (47.9)
74 (39.8)
0.328
CRP > 8 mg/dl, n (%)
30 (62.5)
79 (42.5)
0.015
K > 5 mEq/L, n (%)
12 (25.0)
10 (5.4)
< 0.001
Na < 135 or > 150 mEq/L, n (%) 31 (64.6)
139 (74.7)
0.203
Ca > 10 mEq/L, n (%)
16 (8.6)
0.274
7 (14.6)
Abbreviations: SBP, systolic blood pressure; HR, heart rate; WBC, white blood cell count;
Hb, hemoglobin; ALT, alanine aminotransferase; BUN, blood urea nitrogen; Cr, creatinine;
CRP, C-reactive protein; K, potassium; Na, sodium; Ca, calcium.
25
Table 3: Results of backward stepwise Cox regression analyses: determinants of patients
who died within 7 days
odds ratio
95% CI
p-value
score
no chemotherapy
3.137
1.427-6.896
0.004
1
HR > 120 BPM
2.791
1.244-6.264
0.013
1
WBC > 11000/mm3
3.575
1.585-8.067
0.002
1
Platelet < 130000/mm3
3.657
1.564-8.551
0.003
1
Cr > 1.3 mg/dl
3.295
1.495-7.261
0.003
1
K > 5 mg/dl
3.760
1.237-11.426
0.020
1
Abbreviations: HR, heart rate; WBC, white blood cell count; Cr, creatinine; K, potassium
26
Figure 1: Receiver operating characteristic (ROC) curve of Objective Palliative Prognostic
Score.
Abbreviation: AUC, area under the curve.
27
Figure 2: Mortality rate within 7 days based on Objective Palliative Prognostic Score
28