a b s t r a c t

Objective The lack of a sensitive, practical bedside test for hypovolemia has rekindled interest in the Shock Index (heart rate divided by systolic blood pressure). Here, we compare the effect of Blood donation on standing shock index values with its effect on values for the supine shock index and orthostatic change in shock indicies (OCSI).

Methods: This is a re-analysis of data collected for an earlier report. Data were available from 292 adults below age 65 and 44 adults ages 65 and over, donating 450 mL of blood. We obtained supine and standing vital signs before and after donation and then calculated 95% confidence intervals for differences based on the t-distribution.

Results: Blood donation resulted in a mean increase in the standing shock index of 0.09 [95% CI, 0.08-0.11] in younger adults and 0.08 [95% CI, 0.05-0.11] in older adults. These changes were similar to those noted for OCSI (young, 95% CI, 0.08-0.10; old, 95% CI, 0.04-0.10). Supine shock index values did not change with donation in younger donors (mean difference 0.0 [95% CI, 0.0-0.01]) or older donors (mean difference 0.0 [95% CI, -0.01-

0.03]).

Conclusion: Blood donation does not affect the supine shock index, but it does result in changes in standing shock index that are similar to changes in more complicated orthostatic vital signs.

(C) 2016

Introduction

In 1994, we used a large blood donor cohort to compare criteria for orthostatic vital Sign testing [1]. Over twenty years later, orthostatic vital signs are still considered a useful bedside test for hypovolemia, but they are applied variably in clinical practice, with confusion over their proper measurement and interpretation [2-4].

The shock index, calculated by dividing the heart rate (in beats/min)

by the systolic blood pressure (in mm Hg), was described in 1968 and has been a topic of renewed interest [5]. In 1996, the difference between standing and lying shock indicies, the orthostatic change in shock indicies (OCSI), was described as a single-value measurement of ortho- static changes in vital signs [6]. More recent studies demonstrated that the shock index can be used to predict the need for critical interventions and referrals in hypovolemic pregnant and post-partum women in a re- source-limited setting, to guide therapy in sepsis, and to predict mortal- ity in elderly emergency department patients with co-morbidities [7- 9]. In 2015, a group of researchers requested supine shock index data from our blood donor cohort, prompting a re-analysis of the previous dataset. This re-analysis suggested that the standing shock index, by it- self, might serve as a simple test for hypovolemia. This report compares the effects of a 450-mL acute blood loss on supine shock index with its

E-mail address: [email protected].

effects on the standing shock index and OCSI in a cohort of healthy adult blood donors.

Methods

Data collection and setting

This cohort and the methods of measurement have been described previously [1]. The study population consisted of healthy adult volun- teers donating a standard 450-mL blood volume. Exclusion criteria were anti-hypertensive therapy or history of hypertension. A group se- quential design was used to collect data from 292 donors under the age of 65 and from 44 who were 65 years of age or older. Before and after blood donation, heart rate and blood pressure were measured, using an automated cuff. The cuff was activated in the supine position after a minimum 5-min rest and at 1 min and 2 min after standing. Partici- pants provided oral informed consent. Institutional review board ap- proval was obtained for this study.

Data analysis

Mean and standard deviation values were calculated for supine and standing heart rate (beats/min), systolic blood pressure (sbp, in mm Hg), and shock index. Mean and standard deviation (SD) values were also calculated for paired differences between before-donation and after-donation values for supine and standing heart rate, supine and

http://dx.doi.org/10.1016/j.ajem.2016.12.052

0735-6757/(C) 2016

638 M.D. Witting / American Journal of Emergency Medicine 35 (2017) 637-639

standing spb, supine and standing shock index, and for the OCSI (stand- ing shock index minus supine shock index). These calculations were performed separately for the younger and older groups. For each paired parameter, the coefficient of variation (CV) - SD/mean - was calculated as a measure of variability in effect relative to effect size; parameters with lower CV are better discriminators than those with higher CV. The t-distribution was used to calculate 95% CI for all mean values. All standing parameters were calculated using values obtained after 1 min of standing. Diastolic blood pressure or 2-min standing values were not analyzed because the earlier paper showed no benefit of these measurements [1].

Because the two age groups had similar values for the standing shock index, they were pooled to calculate test characteristics for the standing shock index at a variety of cutpoints. Likelihood ratios and con- fidence intervals were calculated as described by Simel [10]. Cutpoints were selected that were diagnostic, having relatively extreme values for likelihood ratios, sensitivity, specificity, or had a high diagnostic yield, the proportion of tests with diagnostic results.

Results

Table 1 shows the effect of a 450-mL blood loss on vital signs as well as the effects of supine versus standing positioning. Supine pulse, sbp, and shock index values showed little effect of blood donation, as evi- denced by high CV values and confidence intervals for differences span- ning 0 (zero). Standing pulse and sbp values showed clear effects from blood donation, but effects differed in younger and older groups. Of the parameters tested, only standing shock index and OCSI had large ef- fects relative to their variation (low CV) and consistent effects for youn- ger and older groups. The effects of blood donation on OCSI and standing shock index were similar to each other, as measured by mean differences and SD of mean differences.

Table 2 shows test characteristics for a variety of cutpoints for the standing shock index. The vast majority of blood donors had standing shock index values of 0.5 or more after donation (sensitivity 96%), and a cutpoint of 0.6 had a likelihood ratio for a negative test of 0.5. In con- trast, a cutpoint >=1.0 had a specificity of 99%, and a cutpoint >= 0.9 had a likelihood ratio for a positive test of 4.1.

Discussion

This analysis demonstrates that a simple parameter, the standing shock index, is superior to supine vital signs in distinguishing euvolemic patients from those with moderate hypovolemia. The discriminative ability of the standing shock index was similar to that of OCSI, which re- quires supine vital sign measurement and more involved computation. In 1980, Knopp demonstrated the superiority of orthostatic vital signs over supine vital signs in detecting blood loss [11]. A 1999 system- atic review confirmed the superiority of postural changes over supine

Table 2

Test characteristics for various standing shock index cutpoints in detecting a 450-mL blood loss

Cutpoint	Sensitivity	Specificity	LR+ [95% CI]	LR- [95% CI]
>= 0.5	96% (323/336)	9% (30/336)	1.1 [1.0,1.1]	0.4 [0.3,0.7]
>= 0.6	84% (282/336)	35% (116/336)	1.3 [1.2,1.4]	0.5 [0.4,0.6]
>= 0.7	61% (205/336)	61% (205/336)	1.6 [1.3,1.8]	0.6 [0.6,0.7]
>= 0.8	36% (120/336)	88% (295/336)	2.9 [2.1,4.0]	0.7 [0.7,0.8]
>= 0.9	13% (45/336)	97% (325/336)	4.1 [2.2,7.8]	0.9 [0.9,0.9]
>= 1.0	6% (20/316)	99% (331/336)	4.0 [1.5,11]	1 [0.9,1.0]

For clinical utility, High-yield criteria are in bold print; alternate criteria are italicized. LR+, likelihood ratio of a positive test; LR-, likelihood ratio of a negative test.

measurement [2]. Nevertheless, confusion over the performance of or- thostatic vital signs persists. Protocols describe rest periods ranging from 2 to 30 min prior to testing [12]. Even after the rest period, pro- viders differ in their measurement protocol, with some protocols being as short as 1 or 2 min and others as long as 10 min [4]. There is also wide variation in the interpretation of orthostatic vital signs. Among 53 urban ED providers, 15 different specific criteria were used to denote a positive test, and 28 respondents named criteria that were either unclear or matched only one other response. Many criteria for positive orthostatic vital signs include cutpoints for two or more param- eters, along with an “and” or an “or,” yielding many degrees of freedom [4]. Arithmetic combinations, such as OCSI and ratio of orthostatic shock indices, produce a single value but are computationally challenging and therefore not widely used [5,6,13].

The standing shock index offers a simple alternative to orthostatic vital sign measurements. In our data, it had discriminative ability similar to that of the OCSI, a parameter that performs similarly in younger and older adults [6]. Eliminating the need for a supine measurement de- creases time spent and time variability: there is no resting period, time to measure supine vital signs, or time to wait between measure- ments. It is also much easier to compute than the OCSI.

One potential use for the standing shock index is during emergency department triage, arguably the best time to detect hypovolemia. Since supine measurement is impractical at triage, orthostatic changes from sitting to standing have been described [13]. Still, one study indicated that only 57% of nurses obtain orthostatic vital signs at triage [4]. The standing shock index can be measured easily at triage, before an ambu- latory patient sits down.

Based on study data, two high-yield cutpoints for standing shock index are 0.6 and 0.9. A standing shock index b 0.6 offers evidence against hypovolemia (LR- 0.5), while a standing shock index >= 0.9 of- fers evidence in favor of hypovolemia (LR+ 4.1). These likelihood ratios are modest compared with those desired for a diagnostic test (LR- of 0.1 or LR+ 10), but this is a clinical finding, rather than a diagnostic test. Also, based on spectrum bias, one would expect better discrimina- tion in a model of greater blood loss. Alternative cutpoints of 0.5 and 1.0

Table 1

Vital signs before and after donation of 450 mL of blood.

	Supine heart rate	Supine SBP	Supine shock index	Standing shock index	OCSI	Standing heart rate	Standing SBP
Adults b 65 yr (n = 292)
Before donation	70.5 (10.4)	127 (13)	0.56 (0.09)	0.67 (0.13)	0.11 (0.09)	80.2 (12.3)	122 (15.1)
After donation	69.7 (9.8)	125 (13)	0.56 (0.09)	0.76 (0.16)	0.20 (0.12)	88.3 (14.0)	118 (14.4)
Difference	-0.8 (5.4)	-1.9 (8.4)	-0.003 (0.06)	0.09 (0.10)	0.09 (0.11)	8.2 (8.2)	4.0 (10.5)
CV	6.7	4.5	20.8	1.1	1.2	1.0	2.6
[95% CI for diff.]	[-1.4,-0.2]	[-2.8,-0.9]	[0,0.01]	[0.08,0.11]	[0.08,0.10]	[7.2,9.1]	[2.8,5.2]
Adults >= 65 yr (n = 44) Before donation	73.8 (9.9)	138 (14)	0.50 (0.08)	0.61 (0.12)	0.11 (0.07)	77.7 (10.7)	130 (18.3)
After donation	73.8 (9.9)	136 (17)	0.51 (0.09)	0.69 (0.15)	0.18 (0.11)	82.9 (12.5)	123 (18.8)
Difference	-0.1 (5.0)	-2.3 (11)	0.009 (0.06)	0.08 (0.10)	0.07 (0.09)	5.2 (5.7)	7.0 (12.3)
CV	37.0	4.7	6.2	1.2	1.3	1.1	1.8
[95% CI for diff.]	[-1.4,1.7]	[-5.5,1.0]	[-0.01,0.03]	[0.05,0.11]	[0.04,0.10]	[3.5,-6.9]	[3.3,10.8]

Values are shown as means, standard deviations (SD) are shown in parentheses, and 95% confidence interval (CI) for differences (diff.) are shown in brackets. SBP is systolic blood pressure in mm Hg. Pulse is in beats/minute. OCSI is orthostatic change in shock indicies.

M.D. Witting / American Journal of Emergency Medicine 35 (2017) 637-639 639

are easy to remember and calculate and have sensitivity or specificity values exceeding 95%.

Limitations

One limitation of this study is the inability to generalize the results to a population with co-morbidities. We excluded patients with hyperten- sion, so our findings do not apply to patients with uncontrolled hyper- tension or those taking medications that affect the shock index. In applying the standing shock index to patients at triage, one must be aware that the walk to the triage area might affect vital signs in some patients, particularly those with chronic obstructive pulmonary disease or other baseline impairment in exercise tolerance. Our blood donors rested as they answered questions prior to donation, minimizing the af- fects of exertion prior to measurement.

Conclusion

Donation of 450 mL of blood does not increase supine heart rate or shock index, but it increases standing shock index by 0.09 [0.08- 0.011] in younger adults (bage 65) and 0.08 (95% CI, 0.05-0.11) in adults over 65 years of age. Standing shock index is simpler and as dis- criminating as orthostatic vital signs.

Meeting/funding

This work was not presented at a meeting and was not funded.

Conflict of interest

None to report.

Acknowledgments

The author thanks Vagner Cavalcanti de Albuquerque, MD, Lais Hel- ena Navarro, MD, PhD, and Regina El Dib, PhD, for their inquiry that led to the re-analysis of prior data. He thanks Linda Kesselring, MS, ELS, for her copyediting of this manuscript.

References

1. Witting MD, Wears RL, Li S. Defining the positive tilt test: a study of healthy adults with moderate acute blood loss. Ann Emerg Med 1994;23:1320-3.
2. McGee S, Abernathy WB, Simel DL. The rational clinical examination: is this patient hypovolemic? JAMA 1999;281:1022-9.
3. Naccarato M, Leviner S, Proehl J, et al. Emergency nursing resource: orthostatic vital signs. J Emerg Nurs 2012;38:447-53.
4. Witting MD, Hydorn SC. Variation in orthostatic vital sign testing in an urban aca-

   demic emergency department. J Emerg Nurs 2013;39:619-22.

   Allgower M, Burri C. Shock-index. Ger Med Mon 1968;13:14-9.

5. Witting MD, Smithline HA. Orthostatic change in shock index: comparison with tra- ditional tilt test definitions. Acad Emerg Med 1996;3:926-31.
6. El Ayadi AM, Nathan HL, Seed PT, et al. Outcomes in women with hypovolemic

   shock: the role of the shock index. PLoS One 2016;11:1-12.

   Tseng J, Nugent K. Utility of the shock index in patients with sepsis. Am J Med Sci 2015;349:531-5.

7. Kristensen AKB, Holler JG, Hallas J, et al. Is shock index a valid predictor of mortality

   in emergency department patients with hypertension, diabetes, high age, or receipt of beta- or calcium channel blockers? Ann Emerg Med 2016;67:106-13.

   Simel DL, Samsa GP, Matchar DB. Likelihood ratios with confidence: sample size es-

   timation for diagnostic test studies. J Clin Epidemiol 1991;44:763-70.

   Knopp R, Claypool R, Leonardi D. Use of the tilt test in measuring acute blood loss. Ann Emerg Med 1980;9:29-32.

8. Lance R, Link ME, Padua M, Clavell LE, Johnson G, Knebel A. Comparison of different methods of obtaining orthostatic vital signs. Clin Nurs Res 2000;9:479-91.
9. Witting MD, Gallagher K. Unique cutpoints for sitting-to-standing orthostatic vital signs. Am J Emerg Med 2003;21:45-7.