Value of vital signs in the diagnosis of impaired consciousness: A Cross-sectional observational study


Masayuki Ikeda, M.D., Takashi Matsunaga, M.D., Noritsugu Irabu, M.D., and Shohji Yoshida, M.D.

Running Title: Vital signs in impaired consciousness

Codirector, Department of Clinical Research (M.I.), National Saigata Hospital
Director, Department of Neurology ( T.M ) , Asahi General Hospital
Director, Department of Emergency Medicine ( N.I. ), Asahi General Hospital
Director, Department of Medicine ( S.Y. ), Asahi General Hospital

Correspondence and reprint requests to: Dr. Masayuki Ikeda
Department of Clinical Research, National Saigata Hospital
Ohgata-machi, Nakakubiki-gun, Niigata 949-3193, JAPAN
Phone +81-255-34-3131, Fax +81-255-34-6761
(e-mail: massie@saigata-nh.go.jp)

"Brain lesions often are associated with bradycardia and hypertension. Ikeda and colleagues conducted an observational and cross-sectional study on 529 consecutive patients with impaired consciousness and found that systolic blood pressure has discriminatory power in detecting a brain lesion. They suggest that a Riva-Rocci sphygmomanometer helps in the diagnosis of unconsciousness. "


Abstract
Objectives To determine whether the vital signs offer a quick measure to identify a brain lesion in patients with impaired consciousness.
Design Cross-sectional observational study.
Setting An urban medical centre in Japan.
Participants  529 consecutive patients (mean age, 65 years) with impaired consciousness, with a score less than 15 on the Glasgow Coma Scale.
Main outcome measures We quantified the relationship between the vital signs on arrival and the final diagnosis of a brain lesion by means of the receiver operating characteristic curve (ROC). Stratum-specific likelihood ratios  (SSLRs) were calculated to define strata with optimal discriminant power.
Results Of the 529 patients, 312 (59 %) had a brain lesion which account for the impaired consciousness. The ROC area under the curve and its standard error (ROC AUC } SE) of systolic blood pressure was 0.90 } 0.01, indicating significantly higher accuracy in the identification of a brain lesion than those of the diastolic pressure and pulse rate (ROC AUC } SE: 0.82 } 0.02 and 0.63 } 0.03, respectively, P < 0.001). SSLRs of systolic blood pressure lower than 90 mmHg were less than 0.04, whereas those of systolic blood pressure higher than 170 mmHg were more than 6.09.
Conclusions Among the vital signs, systolic blood offers a useful measure in the diagnosis of impaired consciousness.

What is already known on this topic
In the diagnosis of impaired consciousness, both brain imaging studies and detailed neurological examination often result not only in wastes of time and human and economic resources but also a delay in the correct diagnosis.
The changes in the vital signs associated with an acute insult to the brain may offer a quick measure to identify an organic brain lesion.

What this paper adds
Systolic blood pressure can distinguish between patients with impaired consciousness at high and low risk of an organic brain lesion.
General application of systolic blood pressure to the diagnosis of impaired consciousness helps to avoid wastes of diagnostic modalities and to identify the cause.


Introduction
 The diagnosis of impaired consciousness is always a challenging task for physicians. Although cranial computed tomography (CT) visualises brain lesions, it does not answer to the extrinsic or metabolic brain dysfunctions, which make up more than half of the cases with impaired consciousness [1]. As physicians are not willing to accept the risk of missing an abnormality [2], the use of CT to screen patients with impaired consciousness for a brain lesion has become routine. In an emergency room, however, there are many situations in which we have to save time for brain imaging and put other examinations first, e.g. hypoglycaemic coma, drug poisoning and hepatic encephalopathy. Previous studies suggest that the vital signs help.

 First, hypertension is often observed in patients with acute stroke and in those with increased intracranial pressure [3, 4]. The classic 'Cushing response' [5] (bradycardia with hypertension) is a well-recognised clinical manifestation of increased intracranial pressure. On the other hand, hypotension and tachycardia are the usual findings in states of depressed consciousness that are due to metabolic brain dysfunction, e.g., intoxication, endocrine diseases and sepsis [6].

 The continuous scale of the vital signs is relevant to the receiver operating characteristic (ROC) curve [7, 8].  The ROC curves describe the performance of a test by plotting the test's sensitivity against its false-positive rate (1-specificity) at different levels of positivity or negativity [7, 8].  The ROC curve analysis is a preferred method to evaluate screening tests with a continuous scale used to distinguish between subjects with and without the disease.

 The aim of our study was to examine the potential value of the vital signs to distinguish patients with impaired consciousness who are likely to have an intracranial lesion from those who are not. To this aim, we plotted ROC curves for the ability of each vital sign to identify patients with a brain lesion. Cut points were determined by calculating stratum-specific likelihood ratios (SSLRs) [7] for the different levels of the vital signs. The SSLRs with their 95% confidence intervals (CIs) are a useful method of defining strata with high discriminating power.


Methods
Patients
 From January 2000 to December 2000, we conducted a cross-sectional observational study at the Department of Emergency, Asahi General Hospital, Asahi, Chiba, Japan, an urban medical centre that serves primarily a general population with about 15,000 annual admissions. In every patient the attending physician or nursing staff measured the systolic (SBP) and diastolic (DBP) blood pressure, pulse rate (PR), body temperature (BT) and the score on the Glasgow Coma Scale [9] on arrival. Impaired consciousness was defined in patients with a score of less than 15 on the Glasgow Coma Scale.
 We prospectively evaluated 529 consecutive patients with impaired consciousness, who were at least fifteen years old. We excluded patients with head injury because the cause of the impaired consciousness was obviously due to an insult to the brain. All patients with impaired consciousness were admitted to the ward and were followed up until discharge to determine the cause of the impaired consciousness. We used the final diagnosis which accounted for the impaired consciousness as the gold standard.
 We did not use a standardised diagnostic protocol because we wanted to evaluate the utility of the vital signs for identifying patients with a brain lesion in a clinical setting consistent with that in which the test might actually be used. The institutional review board approved the study and waived the requirement to obtain informed consent because we planned to record data ordinarily collected in the evaluation of patients.

Statistical analysis
 We used the ROC curve [7, 8] to qualify the diagnostic performance of the vital signs to identify patients with a brain lesion. The diagnostic accuracy of each vital sign was estimated by calculating the ROC area under the curve (AUC) and its standard error (ROC AUC } SE), by the nonparametric method developed by Hanley and McNeil  [10].  Significance of differences between curves was determined as described by the same authors [10].
 SSLR is generally defined as the ratio of two probabilities, the probability of a given test result when the disease is present, divided by the probability of the same test result when the disease is absent. SSLRs were determined by means of the formula: SSLR= (x1/n1)/(x0/n0), where x1 is the number of subjects with a brain lesion in the stratum, n1 is the total number of subjects with a brain lesion, x0 is the number of subjects who are without a brain lesion in the stratum and n0 is the total number of subjects who are without a brain lesion. The 95% CI was calculated by the logit method 7. All components of the analysis were performed on a personal computer with the formulas previously published 7.


Results
Patient population
 From January 2000 to December 2000, we admitted 15,293 patients of whom 529 (3.46 percent, 235 women) had impaired consciousness of non-traumatic cause. The mean age was 65 years (range, 15 to 97; SD = 21). Table 1 shows the characteristics of the patients according to the presence (n= 312) or absence ( n= 217) of a brain lesion. There was no significant difference between the two groups in age, sex or the Glasgow Coma Scale. Figure 1 shows the age and sex distributions of the two groups. The predominance of the number of women over that of men in the twenties results from cases of attempted suicide with an overdose of psychotropic drugs. Apart from this age group, men constituted the majority under age 70. This age distribution is similar to those of previous studies on patients with impaired consciousness [3], [11].

Vital signs in patients analysed with ROC and SSLRs
Table 2A and Table2B show the cause of impaired consciousness and the vital signs of the patients. Some of the vital signs made a clear distinction between the two groups classified according to the presence of a brain lesion. The group of 312 (59 %) patients with a brain lesion included 259 (49%) patients with stroke, 29 (5.5 %) with epilepsy, 13 (2.5 %) with a brain tumor and 11 (2.1 %) with meningitis or encephalitis. 217 (41%) patients without a brain lesion consisted of 95 (18 %) with diffuse hypoxia or ischaemia, 59 (11 %) with drug poisoning, 20 (3.8 %) with hepatic encephalopathy, 12 (2.3 %) with diabetic coma and 11 (2.1 %) with ionic imbalance. Diffuse hypoxia/ischaemia includes encephalopathy due to hypoxia, hypercapnia or systemic illness such as sepsis. SBP (168 } 36 mmHg vs 111 } 27 mmHg, P < 0.0001) and DBP ( 90 } 19 mmHg vs 67 } 17 mmHg, P < 0.0001) were significantly higher and PR ( 84 } 20 /min vs 94 } 24 /min , P < 0.0001) was significantly fewer in the patients with a brain lesion than those without one. Body temperature made no difference.
 The large variation in the mean of each vital sign, however, indicates that there is considerable overlap between the two groups (Figure 2). That makes it difficult to apply the information to clinical practice. A clearer view of the diagnostic significance of the vital signs can be obtained by determining the sensitivity and specificity of each level. We calculated these values with a two step-approach, ROC AUC and SSLR.
 First, we analysed the diagnostic value of SBP, DBP and PR with ROC AUC } SE.  Figure 3 shows the ROC curves for each vital sign reported on a continuous scale. The analysis resulted in a ROC AUC } SE of 0.90 } 0.01 for SBP, 0.82 } 0.02 for DBP and 0.63 } 0.03 for PR. The curve for SBP occupies the highest position in the graph, signifying the best diagnostic performance of the three vital signs. Accordingly, the ROC AUC was significantly higher for SBP than DBP (z= 4.53, P< 0.01).
 We then calculated SSLRs with their 95% CIs to assess the discriminating power of the strata used to plot each ROC curve (Table 3). For SBP (Table 3A), which showed the greatest diagnostic utility as suggested by the ROC analysis, eleven SSLRs could be identified. With the pretest probability of 0.59 (312 cases with a brain lesion out of total of 529) in our study and Bayesf theorem, we estimated the post-test probability for each stratum. Eight SSLRs for DBP (Table 3B) and ten for PR (Table 3C) were of less diagnostic value than SBP.

( Legends to the Figures )

Discussion
 Patients with impaired consciousness are divided into two groups, those with a brain lesion which accounts for the impaired consciousness and those without  [1]. Our study selected SBP as the best vital sign to discriminate between these two groups. Previous clinical batteries to determine the nature and site of the lesion impairing consciousness define the individual criteria [1], [11], [12]. Their criteria are comprehensive but still too complex to be evaluated for general practice.
 Swets [13] classified the accuracy of ROC AUC values between 0.5 and 0.7 as low, those between 0.7 and 0.9 as moderate, and those above 0.9 as high. Judged by their ROC AUC values, the accuracy of DBP (ROC AUC = 0.82) is moderate and that of SBP (ROC AUC = 0.9), significantly higher than that of DBP, is marginally high.
 Not only the classic Cushing response 5 but also previous studies on systemic response to impaired autoregulation of cerebral blood flow (CBF) explain the increase in systemic blood pressure in patients with impaired consciousness due to a brain lesion. In animal experiments [14], [15], brain ischaemia results in an increase in systemic blood pressure. Up to 84% of cases of stroke show an increase in blood pressure in the acute phase [4]. Yamashiro et al. [3] suggested that hypertension can identify patients who have a high probability of stroke as a cause of impaired consciousness. In contrast to hypertension in patients with an organic brain lesion, hypotension, usually associated with metabolic brain dysfunction due to drug intoxication, sepsis, diabetic coma and hepatic encephalopathy [6],  should make SBP a much better measure to discriminate patients with a brain lesion from those without one.
 When CT is available in an emergency department, clinicians liberally order brain a CT scan for virtually every patient with impaired consciousness. Nevertheless, about half of patients with impaired consciousness both in Plum and Posnerfs series [1] and ours had no organic brain lesion. The question we wished to address was whether some of the patients who are now considered candidates for brain imaging can be classified as having a low probability of a brain lesion on clinical grounds. Our study indicates that SBP has discriminatory power for this purpose.
 The characteristics of the subjects in our study may make it subject to certain limitations to generalisation. First, the older mean age, 65 years, than those in previous studies on patients with impaired consciousness, 58 years [11] and 54 years [3], may result from the aged population in our local community. Second, the incidence of stroke (49%) in our study was higher than those of Plum and Posnerfs (26%: 130 cases out of 500) [1] and Yamashiro and others (28%: 49 out of 175) [3] but lower than 57 % among British patients with coma [11]. Third, the high incidence of diffuse hypoxia or ischaemia (18 %) probably reflects the vulnerability of the older patients in our study to severe systemic infections, for example, pneumonia, pyelonephritis and cholecystitis. Bates et al. [11] observed a similar high incidence of diffuse hypoxia or ischaemia among patients with coma.
 In summary, our study confirms the validity of SBP for identifying patients with impaired consciousness who have a brain lesion and those who do not. General application of SBP to the diagnosis of impaired consciousness should result in both clinical and economic benefit. Our study is an important step toward helping clinicians more effectively manage a common problem. We encourage other practitioners to validate the accuracy and generalisability of our results.

Contributors
Masayuki Ikeda designed the study, analysed the data and wrote the report. Takashi Matsunaga and Noritsugu Irabu did the study and collected the data. Shohji Yoshida was medical superintendent of the hospital and allowed us to do our study there. He was also involved in preparation of both the study and the report. Masayuki Ikeda is guarantor for the study.
Acknowledgments
We gratefully acknowledge the contribution in statistics made by Professor Toshiaki Furukawa.


References
1. Plum F, Posner JB. The Diagnosis of Stupor and Coma. 3rd ed. Philadelphia: F. A. Davis; 1980.
2. Graham ID, Stiell IG, Laupacis A, O'Connor AM, Wells GA. Emergency physicians' attitudes toward and use of clinical decision rules for radiography. Acad Emerg Med 1998;5:134-40.
3. Yamashiro S, Oda Y, Kanegae S, Shirahama M, Yoshihara K, Fukui T, et al. Informative usefulness of age, sex and vital signs in the differential diagnosis of disturbed consciousness among 175 emergency outpatients. Fukuoka Igaku Zasshi 1994;85:353-60.
4. Wallace JD, Levy LL. Blood pressure after stroke. JAMA 1981;246:2177-80.
5. Guyton AC, Hall JE. Nervous regulation of the circulation, and rapid control of arterial pressure. In: Textbook of Medical Physiology. 9th ed. Philadelphia: W. B. Sauders; 1996: 209-220.
6. Victor M, Ropper AH. Coma and related disorders of consciouness. In: Adams and Victor's Principles of Neurology. 7th ed. New York: McGraw-Hill; 2001:366-389.
7. Peirce JC, Cornell RG. Integrating stratum-specific likelihood ratios with the analysis of ROC curves. Med Decis Making 1993;13:141-51.
8. Furukawa T A, Goldberg D P, Rabe Hesketh S, Ustun T B Stratum-specific likelihood ratios of two versions of the General Health Questionnaire. Psychol Med , 31 2001;31:519-29.
9. Teasdale G, Jennett B. Assessment of coma and impaired consciousness. A practical scale. Lancet 1974;2:81-4.
10. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839-43.

11. Bates D, Caronna JJ, Cartlidge N, Knill JR, Levy DE, Shaw DA, et al. A prospective study of nontraumatic coma: methods and results in 310 patients. Ann Neurol 1977;2:211-20.
12. Mills ML, Russo LS, Vines FS, Ross BA. High-yield criteria for urgent cranial computed tomography scans. Ann Emerg Med 1986;15:1167-72.
13. Swets JA. Measuring the accuracy of diagnostic systems. Science 1988;240:1285-93.

14. Levy LL, Wallace JD. Cerebral blood flow regulation. II. Vasodilator mechanisms. Stroke 1977;8:189-93.
15. Levy LL, Wallace JD, Stolwijk JA, Poindexter ER. Cerebral blood flow regulation: vascular resistance adjustments in the circle of Willis. Stroke 1976;7:147-50.


Legends to the figures
Figure 1.
Numbers and ages of the patients. Closed columns represent men. Hatched columns represent women.

Figure 2.
Distribution of systolic blood pressure (SBP) (A), diastolic blood pressure (DBP) (B) and pulse rate (PR) (C) among patients with impaired consciousness with (open column) or without (closed column) a brain lesion. SBP (168 } 36 mmHg vs 111 } 27 mmHg, P < 0.0001) and DBP ( 90 } 19 mmHg vs 67 } 17 mmHg, P < 0.0001) were significantly higher and PR ( 84 } 20 /min vs 94 } 24 /min , P < 0.0001) was significantly fewer in the patients with a brain lesion than those without. There is substantial overlap, however, in the values for the two groups of patients.

Figure 3.
Receiver operating characteristic area under the curve (ROC AUC) for systolic blood pressure (SBP), diastolic blood pressure (DBP) and pulse rate (PR). The curves were plotted with the sensitivity and false-positive rate ( 1- specificity) at different levels of the measurement as shown in Table 3.


To Massie's Profile