Khoi D. Than, MD, Anthony C. Wang, MD
Department of Neurosurgery, University of Michigan, Ann Arbor, Michigan, USA
Journal Club Article: Siddiq F, Chaudhry SA, Tummala RP, Suri MFK, Qureshi AI. Factors and outcomes associated with early and delayed aneurysm treatment in subarachnoid hemorrhage patients in the United States. Neurosurgery. 2012;71:670-678.
SIGNIFICANCE/CONTEXT AND IMPORTANCE OF THE STUDY
In “Factors and outcomes associated with early and delayed aneurysm treatment in subarachnoid hemorrhage patients in the United States,” Siddiq et al. analyze a large national database of patients in order to determine which factors, if any, result in early versus delayed treatment of ruptured intracranial aneurysms. Also examined were differences in patient outcomes between the two groups. The authors chronologically present the progression of aneurysm treatment rationale, to the current status where early securement of the aneurysm is essential to minimize the risk of re-bleeding and to maximize prophylaxis and treatment of vasospasm via hypertension. This study is of immense importance to our profession, in that by determining the predictors of delayed aneurysm treatment, we can ideally work toward mitigating those factors and shortening the time to treatment for all subarachnoid hemorrhage (SAH) patients.
ORIGINALITY OF THE WORK
While many studies have examined differences in patient outcomes based on the timing of intervention, to our knowledge, this study represents the first to also look specifically at the factors associated with such timing. A few studies unmentioned in this article, however, have found differences between patient cohorts that are treated early versus late, which we discuss further in Section IX.
APPROPRIATENESS OF THE STUDY DESIGN OR EXPERIMENTAL APPROACH
In this study, the authors utilize the Nationwide Inpatient Sample (NIS) in order to accrue data from 32,048 patients. Appropriate codes were used in order to find patients with aneurysmal SAH who subsequently underwent open or endovascular treatment of an aneurysm. Numerous demographic and outcome variables were analyzed.
Several flaws are noted in the variables collected from the NIS. First is the definition of a “teaching hospital;” namely, “one with an American Medical Association-approved residency program and either membership in the Council of Teaching Hospitals or a ratio of full-time equivalent interns and residents to beds of 0.25 or higher.” This definition includes many hospitals without an academic neurosurgery program and neurosurgery residents. Thus, no statistically significant difference was found in the time to treatment of SAH patients between teaching and non-teaching hospitals. It would be illustrative to see whether exclusively analyzing such hospitals with neurosurgery residency programs (or perhaps with inclusion of only Level 1 trauma centers) results in a significantly earlier time to treatment for aneurysmal SAH patients. One would expect a greater likelihood for differences would be found.
The second flaw involves reported use of the 3M Health Information Systems All Patient Refined Diagnosis Related Group (3M APR-DRG), which is difficult to interpret. The only neurosurgical article that utilized the 3M APR-DRG cited by the authors was written by the same group of authors. According to the primary reference, the 3M APR-DRG takes into account coded principal and secondary diagnoses, coded procedures, age, sex, and discharge disposition. Thus, the scale is dependent on how conditions are coded and does not take into consideration neurologic outcomes at any point in time, length of hospital stay, etc.
Finally, defining “clinical outcomes based on discharge destination” likely results in inaccurate classifications. In our experience, many patients who have good neurologic outcomes (and good insurance policies) still spend a short amount of time in acute rehabilitation and thus would be placed in the “moderate to severe disability” group. Similarly, terminal patients with diminished neurologic outcomes are frequently discharged home to be cared for by family, and here would be lumped into the “none to minimal disability” group. Judging clinical outcomes based on discharge destination is thus inherently very flawed and the reason why most studies look at neurologic outcome scores after significant periods of time (ie, Glasgow Outcome Score [GOS] at 6 months).
ADEQUACY OF EXPERIMENTAL TECHNIQUES
For a crude assessment of the predictors of early versus late treatment of aneurysmal SAH, use of the NIS is adequate. Some inadequacies were mentioned in the previous section. The primary shortcoming of using the NIS is its inherently retrospective nature. A prospective study, even one that results in only a fraction of the total number of patients reported in this study (such as the International Subarachnoid Aneurysm Trial [ISAT], which prospectively analyzed 2,143 patients), would provide more accurate and neurosurgically pertinent information.
SOUNDNESS OF CONCLUSIONS AND INTERPRETATION
The authors found that the predictors of early intervention included endovascular treatment and female sex. Endovascular capabilities certainly allow for more rapid treatment, as the authors acknowledge. While it was statistically significant that women were more likely to undergo earlier intervention (which the authors attribute to aneurysm location and/or number and social factors), the clinical significance is questionable, as females comprised 69% of the early group and 66% of the delayed group (P = 0.01). Patients in the early group had higher in-hospital mortality but survivors were more likely to be discharged home, which the authors attribute to more aggressive treatment of high-grade patients and improved outcomes, respectively. The trend toward more patients in coma/stupor (Hunt-Hess grades 4-5) in the early group lends support to this theory.
A somewhat surprising finding was the higher rate of ventriculostomy in the early group (34% early vs. 29% late; P = 0.001). One wonders whether the lower rate of ventriculostomy (and hence, monitor/control of elevated intracranial pressure) contributed to poorer outcomes and a higher rate of discharge to extended-care facilities in the delayed treatment group.
Patients included in the NIS were more likely to have delayed treatment of their ruptured aneurysms if they were admitted on the weekend. Although this finding is disturbing, it certainly makes sense, given that it is much more difficult to mobilize the necessary teams (surgical, radiological, anesthetic, neuromonitoring, nursing, technical, etc.) during the weekend. Intuitively, treatment was also delayed if patients had pre-existing hypertension, diabetes mellitus, or renal failure. The authors found that patients in the delayed group were more likely to receive gastrostomy tubes and blood transfusions and had longer lengths of hospital stay and subsequent hospital costs, all factors probably attributed to the reported lower rate of in-hospital mortality and lower rate of discharge to home.
RELEVANCE OF DISCUSSION
In their discussion, Siddiq et al. nicely review the probable explanations behind their findings. They discuss early versus delayed patient outcomes for the major aneurysm trials, namely the Intraoperative Hypothermia for Aneurysm Surgery Trial, the International Cooperative Study on the Timing of Aneurysm Surgery, and the ISAT. They also comprehensively discuss the shortcomings of their study.
CLARITY OF WRITING, STRENGTH AND ORGANIZATION OF THE PAPER
The paper is easy to read and is well-organized.
ECONOMY OF WORDS
The authors demonstrate good practice of economy of words. One area where more description would have benefited the readers is further explanation of the 3M APR-DRG, perhaps by including descriptive scenarios that demonstrate how a patient gets categorized as a 1, 2, 3, or 4. As it stands, assignment of that number is unclear and appears to be based only minimally on patient outcomes.
RELEVANCE, ACCURACY AND COMPLETENESS OF BIBLIOGRAPHY
A number of studies addressing this very topic were absent from the bibliography of this article. In a prospective study of 550 patients who underwent open clipping of ruptured aneurysms, Ross et al.1 divided patients into early (post-bleed day 1-3), intermediate (post-bleed day 4-10), and late (post-bleed day 11-21) groups. The authors found no statistically significant difference in patient demographics, pre-operative clinical condition, GOS, time to discharge, or discharge destination between the 3 groups. Predictors of poor outcome, regardless of treatment group, were older age and World Federation of Neurological Surgeons grade of 4 or 5. Late surgery was attributed to “delays in diagnosis, transfer, and logistic factors, but not clinical decision.”
Using the NIS to analyze 12,023 patients who underwent open aneurysm clipping, Cowan et al.2 compared outcomes in patients treated electively and emergently in hospitals of different volumes (very low, low, high, and very high). For patients treated emergently, very low volume hospitals had higher mortality rates than very high volume hospitals (14.7% vs. 8.9%, respectively; P < 0.001). For all patients, predictors of death included renal disease, older age, chronic obstructive pulmonary disease, and emergency admission. Patients admitted to the highest volume quartile of hospitals were more likely to survive than if admitted to hospitals of any of the other 3 quartiles.
In a multi-center retrospective study of 411 patients who underwent open clipping of ruptured aneurysms, Nieuwkamp et al.3 found no difference in outcome between early (post-bleed day 0-3), intermediate (post-bleed day 4-7), and late (post-bleed day 7+) treatment groups, if the patients were in good clinical condition at admission. Patients who were high-grade did have improved outcomes with earlier intervention. Statistics were not calculated to compare demographic variables, but similar to the Siddiq et al. study, there was a trend toward earlier treatment in females. Women made up 72%, 67%, and 64% of the early, intermediate, and late groups, respectively.
Finally, in a prospective analysis of 439 SAH patients, Eden et al.4 did not find that race was a factor in time to treatment in the greater Cincinnati area. The authors did find, however, that patients transferred from other facilities were treated more quickly than patients directly admitted to the emergency department, which was thought secondary to delayed triage times in higher-volume emergency rooms. Similar to the Siddiq et al. study, this study found that patients with diabetes experienced a delay in time to treatment.
NUMBER AND QUALITY OF FIGURES, TABLES AND ILLUSTRATIONS
The 3 tables included by the authors are easy to understand and nicely summarize the study’s findings.
A multi-center prospective study, with patients followed by neurosurgeons and data values understandable by neurosurgeons, would help to validate the findings of this article.
We would like to thank Holly Wagner for her editorial assistance in preparing this document.
1. Ross N, Hutchinson PJ, Seeley H, Kirkpatrick PJ. Timing of surgery for supratentorial aneurysmal subarachnoid haemorrhage: report of a prospective study. J Neurol Neurosurg Psychiatry. 2002;72(4):480-484.
2. Cowan JA, Jr., Dimick JB, Wainess RM, Upchurch GR, Jr., Thompson BG. Outcomes after cerebral aneurysm clip occlusion in the United States: the need for evidence-based hospital referral. J Neurosurg. 2003;99(6):947-952.
3. Nieuwkamp DJ, de Gans K, Algra A, et al. Timing of aneurysm surgery in subarachnoid haemorrhage--an observational study in The Netherlands. Acta Neurochir (Wien). 2005;147(8):815-821.
4. Eden SV, Morgenstern LB, Sekar P, et al. The role of race in time to treatment after subarachnoid hemorrhage. Neurosurgery. 2007;60(5):837-843.