| | The Hardman index in patients operated on for ruptured abdominal aortic aneurysm: a systematic reviewReceived 6 June 2006; accepted 19 July 2006. BackgroundThe aims of the present study were to (1) analyze preoperative predictors for outcome suggested by Hardman and surgical mortality after open repair and endovascular repair (EVAR) of ruptured abdominal aortic aneurysms (rAAA), and (2) further evaluate the Hardman index in a systematic review. MethodsPatients operated on for rAAA during a 5-year period between 2000 and 2004 were scored according to Hardman—1 point for either age >76 years, loss of consciousness after presentation, hemoglobin <90 g/L, serum creatinine >190 μmol/L or electrocardiographic (ECG) signs of ischemia—with blinded evaluation of ECGs by a specialist in clinical physiology. The results were included in a systematic review of studies evaluating the Hardman index. ResultsIn-hospital mortality after operation was 41% (67/162). There was no difference in in-hospital mortality between open repair (n = 106) and EVAR (n = 56), whereas the Hardman index was associated with operative mortality in our institution and in the systematic review of 970 patients (P < .001). Mortality rate in patients with Hardman index ≥3 was 77% in the pooled analysis. A full data set of all five scoring variables was obtained in 94 (58%) of 162 patients in our study, and potential underscoring was thus possible in 68 patients. Of the available ECGs, 12 (8.7%) of 138 were judged nondiagnostic. Five studies did not state their missing data on ECG and hemoglobin and serum creatinine concentrations, nor did they specify the criteria for ECG ischemia. ConclusionsA strong correlation between the Hardman index and mortality was found. A Hardman index ≥3 cannot be used as an absolute limit for denial of surgery. The utility of the Hardman index seems to be impeded by variability in scoring resulting from missing or nondiagnostic data. The population in the western world is aging. Forecasts on age distribution clearly indicate that the fastest growing age category is people aged ≥65 years. In 2020, these numbers will have increased by 50%, relative to 2001.1 It is therefore not surprising that several large cohort studies have indicated an increase in incidence and mortality, age-standardized or not, of abdominal aortic aneurysms (AAA)2, 3, 4, 5 in the population. Risk group identification for direction of effective prevention and screening are awaited. Hence, predictive tools for selecting patients with ruptured AAA (rAAA) for intervention or for making grounded decisions to withhold treatment have become increasingly important. There is general agreement that decision aids improve knowledge and realistic expectations as well as reduce decisional conflicts and the proportion of patients left without a decision.6 To support clinical judgement, the decision tool needs to be based not only on clinically relevant parameters but also to be feasible to use in the clinical situation. The Hardman index,7 POSSUM (Physiological and Operative Severity Score for the enUmeration of Mortality and morbidity),8 and the Glasgow aneurysm score9 have been suggested to improve selection of patients with rAAA for surgery, based on prediction for outcome. The Hardman index scoring is the simplest and the only system that was originally developed from a cohort of patients undergoing emergency surgery for rAAA. Endovascular aortic aneurysm repair (EVAR) is established in Malmö as the first-line treatment option in patients with AAA10 and also in rAAA,11 despite the requirement of preoperative computed tomographic (CT) scanning. Because EVAR can be performed under local anesthesia, fragile patients who are otherwise unfit for open surgery may tolerate this treatment.12 The aims of the present study were to investigate the relationship between the Hardman index and mortality, overall and when comparing EVAR with open surgery, and to compare the results with those in a systematic review of previously published studies evaluating the Hardman score. Materials and methods  Study population and setting The Department of Vascular Diseases Malmö-Lund is a tertiary referral center with a catchment population of approximately 756,000 inhabitants (2002, Swedish Central Bureau of Statistics, www.scb.se.) in the southernmost part of Sweden. Emergent operations for rAAA are mainly performed in Malmö University Hospital, and EVAR is only performed in Malmö. Thus, patients admitted at hospitals outside of Malmö are generally transferred to Malmö after diagnosis if they are stable. If the patient is considered unstable, the vascular surgeon is transferred to the local hospital. Retrieval of rAAA cases During the study period, all surgical procedures, and all diagnoses assigned to in-patients upon discharge or death were classified according to the International Classification of Diseases, 10th Revision (ICD-10) code and collected in a computerized registry. The identification of patients with the diagnosis of rAAA between January 1, 2000, and December 31, 2004, managed operatively or nonoperatively, was based on ICD code I71.3. Medical records were analyzed. The computerized autopsy registries at Department of Pathology, Malmö University Hospital, and the Institution of Forensic Medicine in Lund, Lund University Hospital, were used to identify the protocols of the patients with rAAA. Definition of rupture, circulatory instability, and operation Rupture was defined as extravasation of blood or hematoma outside the AAA on CT examination, during open repair (OR), or at autopsy. In the absence of autopsy, operation, or CT, a patient with a known AAA with acute symptoms compatible with rupture was also considered to have sustained a rAAA. Patients undergoing emergent surgery for a nonruptured inflammatory or symptomatic AAA were excluded. Circulatory instability was defined as loss of consciousness, either transient or permanent, before operation. Operation was defined as the delivery of an anaesthetic with the intention of performing AAA repair or a note of the start of an operation in the anaesthetic report. The Hardman index Patients undergoing emergent surgery were scored according to Hardman et al.7 The Hardman index is composed of five patient-related preoperative predictors. One point each is given for age >76 years, loss of consciousness after presentation, a concentration of serum creatinine >190 μmol/L, hemoglobin <90 g/L, or acute myocardial ischemia defined as depressed ST segments >1 mm and/or associated T wave changes on electrocardiogram (ECG). The Hardman index may thus vary between 0 to 5 points, and scores of ≥3 points have been associated with 100% mortality. Loss of consciousness after presentation reflects circulatory instability in our study and includes both patients with temporary and permanent unconsciousness before surgery. The ECGs available for interpretation were analyzed by a specialist in clinical physiology (M. D.) who was blinded to other data. All patients who survived an acute operation for rAAA were controlled regarding long-term survival in the Swedish Population Registry. Follow-up mortality data All patients were monitored from the day of surgery until death or January 31, 2006. Mean follow-up time was 23 months (median, 16 months; range, 0 to 72 months). Data on all-cause mortality were retrieved by record linkage with the Swedish Population Registry. Deaths occurring before discharge were counted as in-hospital deaths. The analysis of long-term survival encompassed all deaths occurring between surgery and the end of follow-up. This study was approved by the Research Ethics Committee of the University of Lund. Review of studies evaluating the Hardman index The checklist proposed by the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group 13 was used as a guideline. Two authors (SA and MÖ) independently searched the Medline, Embase, and Cochrane library databases for publications between 1996 and October 2005. A predetermined search strategy was used that included the combination of the medical subject heading (MeSH) terms abdominal aortic aneurysm and rupture, mortality, and Hardman as keywords. The reference lists of retrieved studies were checked for additional relevant articles. For the review, studies evaluating the relationship between numeric value of Hardman score and in-hospital mortality were considered eligible. If needed, supplementary information on number of deaths and patients in respective Hardman score stratum, on median (range) age, and on male–female ratio were requested by correspondence with the lead author of the respective papers and was obtained from three studies.14, 15, 16 Statistical methods Differences in proportions for nominal variables were evaluated using the χ2 or Fisher’s exact test and the Kendall’s τ-b test for ordinal variables. Skewed distributions were expressed in terms of median and range, and the Mann-Whitney U test was used to evaluate differences. The Hardman scoring variables were evaluated in a univariate logistic regression model, and variables that were significantly associated with in-hospital mortality were entered in a multivariate logistic regression model. Survival analysis was calculated using the Kaplan-Meier model, and the log-rank test was used to compare survival rates between strata of the Hardman score. Analysis of correlation between the Hardman score and in-hospital mortality was assessed using the Spearman rank test. Results The Malmö experience The nonoperated on patients with rAAA In all, 85 (34.4%) of 247 patients died without having a surgical procedure, and 34 of these died outside the hospital. In-hospital death without a surgical procedure occurred in 51 patients, of whom 15 were assessed in time but were not considered for surgery. The documented factors (five patients had two factors) that contributed to the nonoperative management of these 15 patients were high age (>80 years) in 13, cancer in 3, suprarenal AAA in 2, chronic obstructive pulmonary disease in 1, and 1 patient refused the procedure. Diagnosis was established at autopsy in 64 patients, after acute CT scan in 8, and on clinical grounds (previous known AAA together with acute symptoms) in 13. Median age was 77 years (range, 54 to 96 years), and 21 (24.7%) were women. Hardman index variables in operated patients The 162 operated on patients, 135 (83%) men and 27 (17%) women, were younger (median age, 74; range, 49 to 89) than the nonoperated on patients (P = .023). There was a significant annual increase in the EVAR/OR ratio between 2000 and 2004 (P < 0.001), reaching up to 16 EVAR vs 15 OR in 2004. Most of the patients were operated on in Malmö (n = 116, 71.6%). In all, 106 (65%) open and 56 (35%) endovascular repairs were performed. Hence, the proportion of EVAR in Malmö was 48% (56/116). Clinical data are presented in Table I. The preoperative median hemoglobin level was 116 g/L (range, 42 to 165 g/L), and 19 patients (16%) had values <90 g/L. The preoperative median serum creatinine level was 120 μmol/L (range, 57 to 508 μmol/L) and 18 patients (16%) had values >190 μmol/L. A preoperative ECG was obtained from 138 patients (85.2%). Two ECGs were not possible to evaluate, one owing to poor quality and one because of pacemaker activity. Myocardial ischemia, according to the definition by Hardman, was seen in 56 (41.2%) of 136 patients. Ten ECGs were nondiagnostic because it was impossible to distinguish myocardial ischemia from nonmyocardial ischemia according to the predefined criteria. A full data set of all five scoring variables was obtained in 94 (58%) of 162 patients. Potential underscoring was thus possible in 68 patients. A total of 114 missing scoring data were found and distributed as follows: One, two, and three missing scores were annotated in 26, 32, and 8 patients, respectively. The variables serum creatinine, hemoglobin, and ECG were missing in 48 (30%), 41 (25%), and 24 (15%) patients, respectively. In-hospital mortality The overall in-hospital mortality after operation was 41% (67/162). A significant association was found between the Hardman index and in-hospital mortality (r = 0.38; P < .001) (Table II), as well as between the Hardman index and the long-term survival rate (P < .001) (Fig). The variables age >76 years, loss of consciousness after presentation, and hemoglobin <90 g/L remained independently associated with in-hospital mortality in the multivariate analysis (Table I). The Hardman index or the in-hospital mortality did not differ significantly between patients undergoing OR or EVAR (Table III). Patients in the EVAR group were older (P = .025), whereas patients in the OR group were more often unconscious after presentation (P = .004) during the study period. In 2004, loss of consciousness after presentation was present in 6 (40%) of 15 and 5 (31%) of 16 patients who underwent OR and EVAR, respectively, (P = 0.61). There was a trend towards a higher Hardman index (P = .063) and in-hospital mortality (P = .098) in patients with a full data set (n = 94) compared with those with an incomplete data set (n = 68). A significant association was found between the Hardman index (P < .001) and score ≥3 (P = .038) with in-hospital mortality for those 94 patients with a full data set. The in-hospital mortality in patients with Hardman score ≥3 and a completed data protocol was 67% (14/21). None of those seven patients in profound shock who developed cardiac arrest requiring preoperative and perioperative active resuscitation survived the operation (five OR, two EVAR). Four patients were judged to have a contained rupture; the three who survived (Hardman index 0, 1, and 1, respectively) were treated with EVAR. Systematic review of studies on Hardman index and mortality Characteristics of included studies The literature search found eight clinical studies,7, 14, 15, 16, 17, 18, 19, 20 one of which had been published in a nonindexed book of abstracts 16 (Table IV). All studies were observational, one of which was prospective. The study by Larzon et al15 and our study presented data after both open and endovascular repair of rAAA. Together with the present study, a total of 970 patients had thus been studied. | | |  | First author | Year⁎ | Country | Inclusion | Study design | Operation type | Rupture definition | Nonoperated/ total cases (%) | Mortality follow-up period | |
|---|
| Hardman7 | 1996 | Australia | 1985-1993 | Retrospective | OR | ND | 21/175 (12) | In-hospital | | | Prance19 | 1999 | UK | 1994-1996 | Retrospective | OR | No | Not stated | 30-day | | | Boyle18 | 2003 | UK/Australia | 2000-2002 | Prospective | OR | No | 21/100(21) | In-hospital | | | Neary17 | 2003 | UK | 1990-2001 | Retrospective | OR | No | 41/232(18) | In-hospital | | | Calderwood14 | 2004 | UK | 1999-2002 | Retrospective | OR | Yes | Not stated | Not stated | | | Tambyraja20 | 2005 | UK | 2000-2001 | Retrospective | OR | Yes | 18/100(18) | Post-op | | | Larzon15 | 2005 | Sweden | 2001-2004 | Retrospective | OR, EVAR | Yes | 9/50(18) | In-hospital | | | Gatt16 | Abs | UK | 1998-2003 | Retrospective | OR | No | 37/96(39) | In-hospital | | | Acosta | 2006 | Sweden | 2000-2004 | Retrospective | OR, EVAR | Yes | 85/247(34) | In-hospital | | | | |
In-hospital mortality in relation to Hardman index A significant correlation between Hardman index value and in-hospital mortality was present in our analysis and in five other studies7, 14, 17, 18, 19; but in three, no statistically significant relationship was found.15, 16, 20 In the pooled analysis, a strong correlation between Hardman index and mortality was found (r = 0.36). As indicated by the r2 of 0.13, the model explained 13% of the variance in mortality. Of the patients with a Hardman index >3, 23% survived the postoperative period. Missing data or data bias In four studies, missing data on preoperative ECG, hemoglobin, and serum creatinine ranged from 6.3% to 27.9%, 13.6% to 25.3%, and 16.2% to 31.6% of the cases, respectively. In five studies14, 15, 16, 17, 20 data on ECG, hemoglobin, serum creatinine, or loss of consciousness were not stated, and criteria for ECG ischemia were not specified (Table II). Another study used a different definition of ECG ischemia than the one originally devised.19 In none but the present study was the validation of ECGs blinded. Discussion The present study on 162 patients showed that the Hardman scoring variables age >76 years, loss of consciousness after presentation, and hemoglobin <90 g/L were independently associated with in-hospital mortality. The variable serum creatinine >190 μmol/L was not found to be a predictor for adverse outcome, which may be attributed to a type II statistical error or to the lack of adjustment to age, gender, and body weight in assessing renal function, or both. The variable ECG ischemia may not be as useful as proposed by Hardman. First, the definition criteria has a poor specificity in the range of 30% to 40 % for detecting myocardial ischemia compared with myocardial contrast echocardiography21 or myocardial perfusion imaging.21, 22 Second, surgeons can not be expected to easily be able to score the ECGs according to definition criteria for myocardial ischemia, as the present study showed that 12 (8.7%) of 138 ECGs were judged as nondiagnostic by a specialist in clinical physiology, mainly due to insufficient discriminating power inherent with ECG graphics. The systematic review further revealed methodologic weaknesses in that most studies did not define the criteria used for ECG ischemia. It also seems that the investigators in all of the other studies interpreted the ECGs while simultaneously aware of other data, which may have contributed to a systematic bias. In the present intention to treat analysis, selection of patients to either OR or EVAR was not supported by a difference in the Hardman index, whereas variable analysis showed that patients within the EVAR group were older, and the OR group had a higher proportion of patients who were unconscious after presentation. However, patients who had loss of consciousness after presentation were equally represented in both groups in the latter part of the study period, and the main reason for selecting patients for OR in 2004 was set by the limitations of emergent EVAR in patients with short proximal infrarenal necks and, to some extent, by narrow iliac arteries. It seems that anatomic unsuitability has become the main obstacle in dedicated EVAR centers.23 On the other hand, these complex aneurysm formations are also more difficult to perform with OR, which might contribute to the relatively high mortality rate in the OR group. Further studies are needed to evaluate whether aneurysm morphology alone is a predictor for outcome. The systematic review on 970 patients showed that the relation between the Hardman index and mortality strengthens the preoperative utility of the Hardman scoring system, at least for some of the variables; however, the lack of availability of data such as ECG and levels of hemoglobin and serum creatinine are of major concern when scoring. In the original study by Hardman et al,7 missing data for individual patients were given the mean value for that variable when scoring, whereas the stated missing data in our study have resulted in an underscoring. The amount of missing data in our study can only partly be explained by the retrospective study design and that the tests simply not were done, since similar incomplete data collection also was found in the only prospective study18 performed. Except for the studies by Hardman et al7 and Prance et al,19 it was not possible to know how missing data were handled in terms of scoring in the other studies. Hence, the calculated mortality of 77% in patients with Hardman score ≥3 may be overestimated owing to underscoring. Conclusion The Hardman scoring variables of age >76 years, loss of consciousness after presentation, and hemoglobin level of <90 g/L predicted outcome in a population where both emergent open and endovascular repair for rAAA were feasible. The use of the ECG as a predictor does not seem appropriate as a decision tool. A Hardman index ≥3 cannot be used as an absolute limit for denial of surgery. The utility of Hardman index seems to be impeded by variability in scoring resulting from missing data. We suggest that it seems to be sufficient to assess age, conscious level, hemoglobin level, and renal function in a prospective protocol on predictors in patients with rAAA. Author contributions Conception and design: SA, MÖ, DB Analysis and interpretation: SA, MÖ, MD, DB, BL Data collection: SA, MD, ZZ Writing the article: SA, MÖ, DB Critical revision of the article: SA, MÖ, DB, BL, MD, ZZ Final approval of the article: SA, MÖ, DB, BL, MD, ZZ Statistical analysis: SA, MÖ Obtained funding: SA Overall responsibility: SA References 1. 1Etzioni D, Liv J, Maggard M, Ko C. The aging population and its impact on the surgery workforce. Ann Surg. 2003;238:170–177. MEDLINE |
CrossRef
2. 2Best V, Price J, Fowkes F. Persistent increase in the incidence of abdominal aortic aneurysm in Scotland, 1981-2000. Br J Surg. 2003;90:1510–1515. MEDLINE |
CrossRef
3. 3Filipovic M, Goldacre M, Roberts S, Yeates D, Duncan M, Cook-Mozaffari P. Trends in mortality and hospital admission rates for abdominal aortic aneurysm in England and Wales, 1979-1999. Br J Surg. 2005;92:968–975. MEDLINE |
CrossRef
4. 4Gillum R. Epidemiology of aortic aneurysm in the United States. J Clin Epidemiol. 1995;48:1289–1298. Abstract |
Full-Text PDF (1177 KB)
|
CrossRef
5. 5Reitsma J, Pleumeekers H, Hoes A, Kleijnen J, de Groot R, Jacobs M, et al. Increasing incidence of aneurysms of the abdominal aorta in The Netherlands. Eur J Vasc Endovasc Surg. 1996;12:446–451. Abstract |
Full-Text PDF (581 KB)
|
CrossRef
6. 6O’Connor AM, Stavey D, Entwistle U, Llewelyn-Thomas H, Rovner D, Holmes-Rovner M, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Library. 2005;. 7. 7Hardman D, Fisher C, Patel M, Neale M, Chambers J, Lane R, et al. Ruptured abdominal aortic aneurysms: who should be offered surgery?. J Vasc Surg. 1996;23:123–129. Abstract |
Full-Text PDF (614 KB)
|
CrossRef
8. 8Copeland GP, Jones D, Walters M. POSSUM: a scoring system for surgical audit. Br J Surg. 1991;78:356–360. 9. 9Samy AK, Murray G, MacBain G. Glasgow aneurysm score. Cardiovasc Surg. 1994;2:41–44. MEDLINE 10. 10Dias N, Ivancev K, Malina M, Resch T, Lindblad B, Sonesson B. Does the wide application of endovascular AAA repair affect the results of open surgery?. Eur J Vasc Endovasc Surg. 2003;26:188–194. Abstract | Full Text |
Full-Text PDF (139 KB)
|
CrossRef
11. 11Resch T, Malina M, Lindblad B, Dias N, Sonesson B, Ivancev K. Endovascular repair of ruptured abdominal aortic aneurysms: Logistics and short-term results. J Endovasc Ther. 2003;10:440–446. MEDLINE |
CrossRef
12. 12EVAR Trial participants. Endovascular aneurysm repair and outcome in patients unfit for open repair of abdominal aortic aneurysm (EVAR trial 2): randomised controlled trial. Lancet. 2005;365:. 13. 13Stroup DF, Berlin J, Morton S, Olkin I, Williamsson D, Rennie D, et al. Meta-analysis of observational studies in epidemiology (A proposal for reporting). JAMA. 2000;283:2008–2012. MEDLINE |
CrossRef
14. 14Calderwood R, Halka T, Haji-Michael P, Welch M. Ruptured abdominal aortic aneurysm (Is it possible to predict outcome). Int Angiol. 2004;23:47–53. MEDLINE 15. 15Larzon T, Lindgren R, Norgren L. Endovascular treatment of ruptured abdominal aortic aneurysms: A shift of the paradigm. J Endovasc Ther. 2005;12:548–555. MEDLINE |
CrossRef
16. 16Gatt M, Martinez M, Perry EP, Barghouti N. Do scoring systems help in predicting survival following ruptured abdominal aortic aneurysm surgery?. Surgeon. 2005;3(Suppl):68. 17. 17Neary WD, Crow P, Foy C, Prytherch D, Heather BP, Earnshaw JJ. Comparison of POSSUM scoring and the Hardman index in selection of patients for repair of ruptured abdominal aortic aneurysm. Br J Surg. 2003;90:421–425. MEDLINE |
CrossRef
18. 18Boyle JR, Gibbs PJ, Shearman CP, Raptis S, Philips MJ. Predicting outcome in ruptured abdominal aortic aneurysm: A prospective study of 100 consecutive cases. Eur J vasc Endovasc Surg. 2003;26:607–611. Abstract | Full Text |
Full-Text PDF (126 KB)
|
CrossRef
19. 19Prance SE, Wilson YG, Cosgrove CM, Walker AJ, Wilkins DC, Ashley S. Ruptured abdominal aortic aneurysms: Selecting patients for surgery. Eur J Vasc Endovasc Surg. 1999;17:129–132. Abstract |
Full-Text PDF (147 KB)
|
CrossRef
20. 20Tambyraja A, Fraser C, Murie J, Chalmers T. Validity of the Glasgow Aneurysm Score and the Hardman Index in predicting outcome after ruptured abdominal aortic aneurysm repair. Br J Surg. 2005;92:570–573. MEDLINE |
CrossRef
21. 21Kaul S, Senior R, Firschke C, Wang XQ, Lindner J, Villanueva FS, et al. Incremental value of cardiac imaging in patients presenting to the emergency department with chest pain and without ST-segment elevation: a multicenter study. Am Heart J. 2004;148:129–136. Abstract | Full Text |
Full-Text PDF (237 KB)
|
CrossRef
22. 22Korosoglou G, Labadze N, Hansen A, Selter C, Giannitis E, Katus H, et al. Usefulness of real-time myocardial perfusion imaging in the evaluation of patients with first time chest pain. Am J Cardiol. 2004;94:1225–1231. Abstract | Full Text |
Full-Text PDF (417 KB)
|
CrossRef
23. 23Peppelenbosch N, Geelkerken R, Soong C, Cao P, Steinmerz O, Teijink J, et al. Endograft treatment of ruptured abdominal aortic aneurysms using the Talent aortouniiliac system: An international study. J Vasc Surg. 2006;43:1111–1122. Abstract | Full Text |
Full-Text PDF (280 KB)
|
CrossRef
a Department of Vascular Diseases, Malmö University Hospital, Malmö, Sweden c Department of Clinical Physiology, Malmö University Hospital, Malmö, Sweden b Department of Vascular Surgery, Uppsala University Hospital, Uppsala, Sweden.  Reprint requests: Stefan Acosta, MD, PhD, Department of Vascular Diseases, Malmö University Hospital, S205 02 Malmö, Sweden.
Competition of interest: none. PII: S0741-5214(06)01366-8 doi:10.1016/j.jvs.2006.07.041 © 2006 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved. | |
|
FULL TEXT01366-8/journalimage?img=PIIS0741521406013668.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0741-5214&ishighres=false&allhighres=false&free=yes','journalimage');)
