Associations between Patient Characteristics and a New,Early Do-Not-Attempt Resuscitation Order after Intracerebral Hemorrhage

Abstract

Background:

Decisions to limit care, including use of a do-not-resuscitate (DNR) order, are associated with increased risk of death after intracerebral hemorrhage (ICH). Given the value that patient surrogates place on the physician's perception of prognosis, understanding prognostic indicators that influence clinical judgment of outcomes is critical.

Objective:

The purpose of this study was to understand the patient variables and comorbid illnesses associated with DNR orders placed on patients within 72 hours after ICH.

Design:

Single-center, retrospective review of medical records of 198 consecutive patients with an admission diagnosis of primary supratentorial ICH between July 2007 and December 2010.

Subjects:

Patients who did not experience a DNR order placement during their primary admission for ICH (non-DNR group) were compared to patients who received a new DNR order in the first 72 hours of admission (DNR group).

Measurements:

Results:

Multiple differences in patient and hospital factors were found between patients receiving a new, early DNR order and those who did not receive a DNR order after ICH. In regression modeling, Caucasian race, direct admission, and higher ICH score were associated with placement of a new DNR order early in the course of injury.

Conclusions:

Race, transfer procedures, and injury severity may be important factors associated with placement of new, early DNR orders in patients after ICH.

Introduction

Despite aggressive therapies, intracerebral hemorrhage (ICH) continues to have a short-term mortality of 30–60% and 10-year survival of only 24%.¹ Only 20% of patients recover independence at six months.² The most frequent direct cause of death following ICH remains withdrawal of life-sustaining interventions, and decisions to limit care, including use of a do-not-resuscitate (DNR) order, are associated with increased risk of death.³

Decisions to limit care often fall to surrogate decision makers in patients with ICH. When considering a decision to limit care, surrogates value complete and honest communication about their loved one's illness and need prognostic information from the medical team to carry out their role as substituted decision makers.⁴ When surrogates choose a limitation to aggressive care such as a new DNR order, this may reflect an expectation of a perceived poor short-term prognosis by medical providers.⁵

Interestingly, although up to 88% of medical decision makers express doubt about a physician's ability to prognosticate,⁶ they continue to rate prognostic information provided by physicians as important.⁷ However, in severe ICH, physicians show variability in their subjective clinical judgment and treatment recommendations.⁸ While prognostic tools such as the ICH score can be considered when predicting outcome, the early subjective clinical judgment of physicians within 24 hours of patient admission may be more closely related to prognosis.⁹

Given the value that patient surrogates place on the physician's perception of prognosis, it is important to better understand what prognostic indicators may influence a physician's subjective clinical judgment of outcomes. Few studies assess what patient characteristics outside of the ICH score a physician may consider to provide guidance about the appropriateness of a new, early DNR order. A fuller understanding of what patient characteristics influence a clinician's subjective judgment may help ensure less variability in prognosis and treatment decisions.¹⁰ The purpose of this study was to understand the patient variables and comorbid illnesses associated with DNR orders placed on patients within 72 hours after ICH.

Methods

Our study was conducted at Duke University Hospital (DUH) Neuroscience Intensive Care Unit. Institutional review board approval was obtained from the Duke University School of Medicine Institutional Review Board before data collection. We retrospectively reviewed medical records of 198 consecutive patients with an admission diagnosis of primary supratentorial ICH between July 2007 and December 2010. Patients with preexisting DNR orders before admission or DNR orders placed greater than 72 hours after ICH onset were excluded. Patients who did not experience a DNR order placement during their primary admission for ICH (non-DNR group) were compared to patients who received a new DNR order in the first 72 hours of admission (DNR group). Patients were considered to have a preexisting DNR order if a scanned inpatient DNR order was found in the patient's chart dated before the day of admission or if an advanced directive indicated the patient's desire to not be resuscitated.

Patient characteristics obtained include demographic data, past medical history, clinical data pertaining to the admission for the ICH, and radiographic images. Demographic data collected included patient age, sex, race, and place of residence before admission. Past medical history obtained included dialysis-dependent renal failure, stroke, dementia, atrial fibrillation, cardiovascular diseases, diabetes, and substance abuse. Characteristics of the injury during the first three days of admission included direct admission or transfer from another facility, systolic blood pressure on admission to the ICU, need for mechanical ventilation, need for ventriculostomy, and occurrence of seizure. The ICH score and all individual components of the score were collected.¹¹

All radiographic images were obtained to determine ICH volumes. Hematoma volumes were determined by manual segmentation of de-identified computed tomography (CT) data sets. Analysis was performed using DICOM formatted images loaded into segmentation software (Osirix version 4.0; Pixmeo, Geneva, Switzerland). An image analyst (C.E.H.) reviewed all images for technical adequacy and performed the initial segmentations. The segmented volumes were then sent to a board certified neuroradiologist (P.G.K.) for further editing and ultimate approval.

Inter-extractor reliability was performed by two persons (J.M. and C.E.H.) counterchecking accuracy during the data collection. Descriptive statistics were used to summarize the patient's demographics and prognostic indicators. Two sample t test and Wilcoxon rank sum test (for continuous variables), chi-squared test, and Fisher's exact test (for categorical variables) were used to examine any differences among the patients by DNR decision-making patterns as appropriate. Statistical significance is considered if p < 0.05. Univariable and multivariable logistic regression model were used to detect factors associated with early DNR. The final multivariable logistic regression model was selected using a reverse stepwise selection method.

Results

Of the 198 patients with supratentorial primary ICH, 9 (4.5%) had a preexisting DNR order and 32 had a DNR order placed greater than 72 hours from admission; thus, they were excluded from review. One hundred and fifty-seven patients were included in the final analysis. Twenty-nine (18.4%) of the 157 patients received a new DNR order during the first 72 hours of admission.

Demographic characteristics are shown in Table 1. Of the non-Caucasian subjects, 74 were identified as African American, 3 as Hispanic, 2 as Asian, and 3 as Native American. Significantly greater percentage of new, early DNR orders were placed on patients older than 80 years, who were Caucasian, admitted from a skilled nursing facilities, directly admitted to Duke Hospital, and with history of relevant comorbidities, compared to patients without a DNR order. No significant differences were found for patients who were dialysis dependent or had a history of stroke or dementia. When considered individually, each component of the ICH score occurred in a significantly greater percentage of patients with new, early DNR orders, compared to those without DNR orders.

Table 1.

Selected Variables from a Single-Center Cohort, Dichotomized by Placement of a New, Early Do-Not-Resuscitate Order after Intracerebral Hemorrhage Onset

Characteristics	Overall (n = 157)	Non-DNR (n = 128)	Early DNR (n = 29)	p
Age, years (SD)	61.9 (15.5)	58.9 (14.7)	75.4 (11.0)	<0.01^#
Age >80 years, n (%)	25 (15.9)	13 (10.2)	12 (41.4)	<0.01
Female, n (%)	79 (50.3)	62 (48.4)	17 (58.6)	0.32
Caucasian, n (%)	75 (47.8)	56 (43.8)	19 (65.5)	0.03
Nursing home resident, n (%)	6 (3.9)	3 (2.3)	3 (11.1)	0.07^*
Direct admission, n (%)	70 (44.6)	52 (40.6)	18 (62.1)	0.04
History of malignancy, n (%)	16 (10.2)	10 (7.8)	6 (20.7)	0.04
History of atrial fibrillation, n (%)	14 (9.0)	8 (6.3)	6 (20.7)	0.01
History of warfarin use, n (%)	19 (12.2)	12 (9.4)	7 (24.1)	0.03
History of hypertension, n (%)	124 (79.0)	99 (77.3)	25 (86.2)	0.45^*
History of antihypertensive, n (%)	84 (53.9)	63 (49.6)	21 (72.4)	0.03
History of dialysis dependent, n (%)	4 (2.6)	3 (2.3)	1 (3.4)	0.56^*
History of prior stroke, n (%)	22 (14.1)	17 (13.4)	5 (17.2)	0.59
History of dementia, n (%)	5 (3.2)	3 (2.3)	2 (6.9)	0.23^*
History of tobacco use, n (%)	35 (22.4)	33 (26.0)	2 (6.9)	0.03^*
History of drug abuse, n (%)	15 (10.3)	15 (12.6)	0 (0)	0.07^*
History of alcohol abuse, n (%)	27 (18.6)	26 (21.8)	1 (3.8)	0.048^*
Systolic blood pressure, mmHg (SD)	176.4 (38.4)	172.8 (38.3)	193.1 (34.9)	0.01
Elevated INR >1.4, n (%)	22 (14.0)	14 (10.9)	8 (27.6)	0.02
ICH score (SD)	1.6 (1.2)	1.3 (1.0)	3.2 (0.7)	<0.01^#
Glasgow coma score, n (%)
13–15	84 (53.5)	82 (64.1)	2 (6.9)
5–12	54 (34.4)	39 (30.5)	15 (51.7)
3–4	19 (12.1)	7 (5.5)	12 (41.4)	<0.01
Interventricular blood, n (%)	67 (43.0)	50 (39.1)	17 (60.7)	0.04
Hematoma volume >30cc, n (%)	55 (37.2)	34 (27.4)	21 (87.5)	<0.01^*

p-values are determined by 2 sample t test or Wilcoxon rank sum test for continuous variables.

^*p-values are determined by chi-squared test or Fisher's exact test for categorical variables.

DNR, do not resuscitate; ICH, intracerebral hemorrhage; INR, international normalized ratio; SD, standard deviation.

In univariable logistic regression analysis (Table 2), several variables were found to be individually significantly associated with placement of a new, early DNR order. Multivariable logistic regression of the variables found to be significant in univariate analysis showed that Caucasian race, direct admission, and the composite ICH score were jointly significantly associated with a new, early DNR order by multivariable logistic regression model (Table 3).

Table 2.

Univariate Analyses of Associations with Placement of a New, Early Do-Not-Resuscitate Order after Intracerebral Hemorrhage (n = 131)

Variable	OR	95% LL OR	95% UL OR	p
Age	1.092	1.049	1.138	<0.001
Age >80 years	4.543	1.672	12.344	<0.001
Female	2.107	0.862	5.148	0.102
Caucasian	2.18	0.892	5.328	0.087
Nursing home resident	7.909	1.248	50.116	0.028
Direct admission vs. transfer	2.992	1.197	7.483	0.019
History of malignancy	3.825	1.221	11.984	0.021
History of atrial fibrillation	4.342	1.354	13.922	0.014
History of warfarin use	3.411	1.109	10.489	0.032
History of hypertension	2.176	0.605	7.829	0.234
History of antihypertensive use	3.627	1.352	9.727	0.011
History of dialysis dependent	1.597	0.159	16.019	0.691
History of prior stroke	1.717	0.56	5.26	0.344
History of dementia	5.044	0.676	37.656	0.115
History of tobacco use	0.224	0.05	1.004	0.051
History of drug abuse	<0.001	<0.001	>999.999	0.962
History of alcohol abuse	0.147	0.019	1.142	0.067
ICH score	8.37	3.568	19.633	<0.001

LL, lower limit; OR, odds ratio; UL, upper limit.

Bolded values have significance of p-value less than 0.05.

Table 3.

Multivariate Analysis of Included Variables and Placement of a New, Early Do-Not-Resuscitate Order after Intracerebral Hemorrhage (n = 131)

Variable	OR	95% LL OR	95% UL OR	p
Caucasian	9.21	1.73	49.019	0.009
Direct admission vs. transfer	5.28	1.06	26.308	0.042
ICH score	11.043	3.901	31.265	<0.001

Discussion

In this single-center cohort of patients with ICH, differences were found between patients receiving a new, early DNR order and those who did not receive a DNR order after ICH. In regression modeling, Caucasian race, direct admission, and higher ICH score were associated with placement of a new DNR order early in the course of injury. Interestingly, the existence of a preexisting neurologic illness, including stroke or dementia, did not associate with the presence of a new DNR order.

Consistent with prior studies, elements of the ICH score (age and intraventricular hemorrhage) have been tied to early DNR order placement. Both ICH score variables (age and hematoma volume) and early DNR order placement are predictive of one-month fatality.¹² In a previous study of ICH patients, those who were made DNR within the first 24 hours were 2.6 times more likely to die than those who were not given a DNR order.¹³ Although likely to be reflective of initial injury severity, implementation of a DNR order has appropriately lead to the concern that early decision to limit care may lead to a self-fulfilling prophecy.³ The American Heart Association recommendations for palliative care in stroke recognized this possible bias and cautioned clinicians to be aware of the inherent uncertainty and limitations surrounding prognostic estimates based on either clinician experience or prognostic models.¹⁴ Despite these recommendations, in this study, the association between new, early DNR order and ICH score may be related to the preconceived expectation of outcome.¹⁵

Direct admission, as opposed to the transfer from an outside facility, was also associated with early DNR in this cohort. While the retrospective nature of the dataset does not allow direct assessment of this relationship, patients with severe ICH injury or poor premorbid function may not be triaged to tertiary centers, since outcome may be deemed too poor. Similarly, due to increasing use of “tele-stroke” services within Duke's network, the transfer process of patients with ICH may represent recognition that more aggressive care at another tertiary facility would be appropriate to achieve a better outcome. Therefore, the decision to transfer may well reflect an escalation of care that runs counter to the intent of a new DNR order in many patients. Future work should seek to determine whether such subtle between-group differences may affect DNR order placement.

Our study did not find a difference between men and women receiving early DNR orders. This is in contrast to prior works where women with ICH were more likely to receive early DNR orders regardless of age or severity scores.^16,17 However, no gender or gender–age interactions were found for mortality or modified Rankin Scale score at 90 days after ICH when evaluated in a multivariate model, but this may be in contrast to other studies.¹⁸ In contrast, critically ill women have less aggressive treatment preferences or are more likely to have surrogates who desire less aggressive treatments than men.¹⁹ At present, the effect of gender on early DNR order placement after ICH remains unclear.

Race/ethnicity and its relationship to end-of-life decisions in ICH are not well studied. Black, Hispanic, and Asian patients with ICH appear to have lower risk-adjusted in-hospital mortality than white patients with ICH,²⁰ but the contribution of new, early DNR orders and/or initiation of comfort measures to this mortality disparity is less clear. For example, less use of comfort measures only for Native Hawaiians and other Pacific Islanders after ICH, compared to whites and Asians, was largely driven by younger age of Native Hawaiians, rather than from underlying cultural differences/ethnicity.²¹ In addition, the likelihood of Mexican-Americans to have DNR orders after ICH, compared to non-Hispanic whites, was attenuated after adjustment for age and other confounders.²² Findings from this cohort add to this growing body of literature, suggesting that racial/ethnic differences may exist, and the relationship between the initiation of comfort measures and mortality is complicated.

While this study did not find associations between comorbidities, premorbid function, and early DNR order placement in univariate or multivariate analyses, prior work has shown a relationship. Bacchetta et al.²³ speculated that surrogate decision makers of patients suffering from the burden of chronic disease may have a more cautious understanding of prognosis. Differences between present and prior findings regarding the influence of comorbidities on DNR order placement are unclear. Use of overall health grading scales, integrating a patient's comorbidities and their severity, may be useful in resolving this issue. Furthermore, larger ICH cohorts would allow for inspection of relative contributions of individual comorbidities to new DNR order placement.

There are several limitations to this study. While this modestly large cohort allowed testing of associations, a larger, prospectively collected sample in future study would provide adequate power to perform a correlative multivariable analyses. Assessing the influence of individual care providers in implementing a new, early DNR order was not assessed in this study, although this factor clearly plays an important role in end-of-life decision making. Information regarding patients' functional status and quality of life was not available and likely plays a role in surrogates' decisions to implement a DNR order.

Surrogates consider prognosis from the physician team when determining limits to treatments, but this is not the only factor that influences decision making.²⁴ To best understand how decisions are made to limit care in patients with ICH, a study that examines factors important to both physicians and surrogate decision makers should be completed. An understanding of these factors may lead to ICH-specific decision aids that improve shared decision making between clinicians and surrogates, decrease variability in prognosis, and emphasize the goals and values of patients.²⁵

Conclusion

Our study identifies differences between ICH patients with a new, early DNR order and patients who receive more aggressive care. Limiting care is the primary cause of death for ICH patients, and careful attention should continue to be paid toward prognostic factors considered by physicians taking care of these patients. Future work should examine how ICH score variables might be used by caregivers to place new DNR orders after ICH.

Footnotes

Acknowledgments

We would like to acknowledge the tireless efforts of the nursing staff in the Duke University Hospital Neuroscience ICU in providing excellent care for patients, and Kathy Gage for her assistance in medical editing. This work was funded by the American Heart Association, Scientist Development Grant (M.L.J.).

Author Disclosure Statement

No competing financial interests exist.

References

Sacco

, Marini

, Toni

, et al.: Incidence and 10-year survival of intracerebral hemorrhage in a population-based registry. Stroke, 2009; 40:394–399.

Manno

: Update on intracerebral hemorrhage. Continuum (Minneap Minn), 2012; 18:598–610.

Becker

, Baxter

, Cohen

, et al.: Withdrawal of support in intracerebral hemorrhage may lead to self-fulfilling prophecies. Neurology, 2001; 56:766–772.

LeClaire

, Oakes

, Weinert

: Communication of prognostic information for critically ill patients. Chest, 2005; 128:1728–1735.

Hakim

, Teno

, Harrell

Jr. , et al.: Factors associated with do-not-resuscitate orders: Patients' preferences, prognoses, and physicians' judgments. SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment. Ann Intern Med, 1996; 125:284–293.

Zier

, Burack

, Micco

, et al.: Doubt and belief in physicians' ability to prognosticate during critical illness: The perspective of surrogate decision makers. Crit Care Med, 2008; 36:2341–2347.

Evans

, Boyd

, Malvar

, et al.: Surrogate decision-makers' perspectives on discussing prognosis in the face of uncertainty. Am J Respir Crit Care Med, 2009; 179:48–53.

Zahuranec

, Fagerlin

, Sanchez

, et al.: Variability in physician prognosis and recommendations after intracerebral hemorrhage. Neurology, 2016; 86:1864–1871.

Hwang

, Dell

, Sparks

, et al.: Clinician judgment vs formal scales for predicting intracerebral hemorrhage outcomes. Neurology, 2016; 86:126–133.

10.

Silvennoinen

, Meretoja

, Strbian

, et al.: Do-not-resuscitate (DNR) orders in patients with intracerebral hemorrhage. Int J Stroke, 2014; 9:53–58.

11.

Hemphill

3rd , Bonovich

, Besmertis

, et al.: The ICH score: A simple, reliable grading scale for intracerebral hemorrhage. Stroke, 2001; 32:891–897.

12.

Brizzi

, Abul-Kasim

, Jalakas

, et al.: Early do-not-resuscitate orders in intracerebral haemorrhage; frequency and predictive value for death and functional outcome. A retrospective cohort study. Scand J Trauma Resusc Emerg Med, 2012; 20:36.

13.

Hemphill

3rd , Newman

, Zhao

, Johnston

: Hospital usage of early do-not-resuscitate orders and outcome after intracerebral hemorrhage. Stroke, 2004; 35:1130–1134.

14.

Holloway

, Arnold

, Creutzfeldt

, et al.: Palliative and end-of-life care in stroke: A statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke, 2014; 45:1887–1916.

15.

Hemphill

3rd , Farrant

, Neill

Jr .: Prospective validation of the ICH score for 12-month functional outcome. Neurology, 2009; 73:1088–1094.

16.

Nakagawa

, Vento

, Seto

, et al.: Sex differences in the use of early do-not-resuscitate orders after intracerebral hemorrhage. Stroke, 2013; 44:3229–3231.

17.

Patel

, Mahajan

, Benjo

, et al.: A national perspective of do-not-resuscitate order utilization predictors in intracerebral hemorrhage. Neurohospitalist, 2016; 6:7–10.

18.

James

, Langefeld

, Sekar

, et al.: Assessment of the interaction of age and sex on 90-day outcome after intracerebral hemorrhage. Neurology, 2017; 89:1011–1019.

19.

Quill

, Ratcliffe

, Harhay

, Halpern

: Variation in decisions to forgo life-sustaining therapies in US ICUs. Chest, 2014; 146:573–582.

20.

Xian

, Holloway

, Smith

, et al.: Racial/Ethnic differences in process of care and outcomes among patients hospitalized with intracerebral hemorrhage. Stroke, 2014; 45:3243–3250.

21.

Shaw

, Gallek

, Sheppard

, et al.: Ethnic differences in withdrawal of life support after intracerebral hemorrhage. Hawaii J Med Public Health, 2015; 74:203–209.

22.

Zahuranec

, Brown

, Lisabeth

, et al.: Ethnic differences in do-not-resuscitate orders after intracerebral hemorrhage. Crit Care Med, 2009; 37:2807–2811.

23.

Bacchetta

, Eachempati

, Fins

, et al.: Factors influencing DNR decision-making in a surgical ICU. J Am Coll Surg, 2006; 202:995–1000.

24.

Boyd

, Lo

, Evans

, et al.: “It's not just what the doctor tells me”: Factors that influence surrogate decision-makers' perceptions of prognosis. Crit Care Med, 2010; 38:1270–1275.

25.

Muehlschlegel

, Shutter

, Col

, Goldberg

: Shared decision making in neurocritical care: An unmet need in our NeuroICUs. Neurocrit Care, 2015; 23:127–130.