Skip Navigation



Annals of Oncology Advance Access published online on May 19, 2007
Annals of Oncology, doi:10.1093/annonc/mdm138
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
18/8/1395    most recent
mdm138v1
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Kelly, L
Right arrow Articles by Stone, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kelly, L
Right arrow Articles by Stone, P.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 2007 European Society for Medical Oncology

The B12/CRP index as a simple prognostic indicator in patients with advanced cancer: a confirmatory study

L Kelly1,*, S White2 and PC Stone2

1 Palliative Medicine, Surrey and Sussex Healthcare NHS Trust
2 Division of Mental Health, St George's University of London, UK

* Correspondence to: Dr L. Kelly, Surrey and Sussex Healthcare NHS Trust, Canada Avenue, Redhill, Surrey RH1 5RH, UK. Tel: +44-1737-768511 ext. 2660; Fax: +44-1737-231949; E-mail: laura.kelly{at}sash.nhs.uk


   Abstract
Top
 Abstract
introduction
 patients and methods
 results
 discussion
 Acknowledgements
 References
 
Background: The vitamin B12/C-reactive protein Index (BCI) has been proposed as a prognostic indicator in patients with advanced cancer. The purpose of this study was to confirm the utility of the BCI in palliative care patients.

Patients and methods: Patients with advanced cancer provided a blood specimen for analysis. Demographic and disease-related variables were recorded. Patients were followed up for at least 90 days or until death.

Results: Patients (n = 329) were divided into three groups according to their BCI score. Patients in group 3 (BCI >40 000; median survival 29 days) had a significantly (P < 0.01) worse survival than patients in group 2 (BCI 10 001–40 000; median survival 43 days) and patients in group 1 (BCI ≤10 000; median survival 71 days). However, patients in group 1 did not have a significantly better prognosis than those in group 2 (P = 0.091). The point estimates for 90-day mortality for each of the three risk groups were different from the figures previously reported during the development phase of the BCI (group 1, 58.9% versus 47.2%; group 2, 64.0 versus 72.5%; group 3, 78.9% versus 90.6%).

Conclusions: An elevated BCI (>40 000) predicts poor survival in patients with advanced cancer.

C-reactive protein, neoplasms, palliative care, prognosis, survival, vitamin B12


   introduction
Top
 Abstract
 introduction
patients and methods
 results
 discussion
 Acknowledgements
 References
 
Health care professionals are unreliable at estimating survival in terminally ill cancer patients [1, 2]. Clinical predictions are usually over-optimistic [2]. It is important for both health care professionals and patients to have realistic expectations about survival so that appropriate management plans can be made. The inability to predict survival accurately may have detrimental effects on patient care. There is evidence that patients who are over-optimistic about their prognosis may opt for futile, aggressive treatments rather than (possibly more appropriate) ‘palliative care’ [3]. Patients themselves regard obtaining prognostic information as of the highest importance [4, 5]. Prognostic information helps patients to make the best use of the time that remains and may allow them the opportunity to make arrangements for their families [6]. Qualitative research has indicated that both patients and their families may regret being over-optimistic about their chances of survival [7]. Moreover, it has been indicated that over-optimism about survival by doctors may contribute to late referral to palliative care services [8]. Better prognostic information may also help with future care planning (for instance deciding whether a patient should be referred for nursing home or hospice inpatient care).

The European Association of Palliative Care (EAPC) have recently published evidence-based clinical recommendations about prognostic factors in patients with advanced cancer [9]. Among six key recommendations, the EAPC advised that clinical predictions of survival should be supplemented by consideration of other prognostic factors or the use of prognostic scales. However, it was noted that only two scales had undergone adequate validation by being developed in one population (the ‘training set’) and then being subject to confirmatory analysis [10] in an independent cohort (the ‘testing set’).

Geissbuhler and co-workers [11] have previously reported the development of the vitamin B12/C-reactive protein index (BCI). They undertook an exploratory study involving a training set of 161 consecutive cancer patients admitted to a geriatric/palliative care unit. All patients had a blood sample taken on the day after admission and a number of haematological and biochemical variables were analysed. Multiple regression analysis identified two variables that were independently predictive of survival [serum vitamin B12 and C-reactive protein (CRP)]. The BCI index was obtained by multiplying the serum vitamin B12 level (pmol/l) by the serum CRP level (mg/l). The BCI index divided patients into three prognostic groups. Group I (BCI ≤10 000) had a 3-month survival of 52.8%, group II (BCI 10 001–40 000) had a 3-month survival of 27.5% and group III (BCI >40 000) had a 3-month survival of 9.4%. However, no confirmatory study of the BCI index was carried out. We therefore decided to test the validity of the BCI in an independent population of patients with advanced cancer.


   patients and methods
Top
 Abstract
 introduction
 patients and methods
results
 discussion
 Acknowledgements
 References
 
This was a prospective, observational multicentre study. The study was approved by the Trent Multi Centre Research Ethics Committee. Patients were recruited prospectively from 21st January 2002 to 8th March 2005 and were followed up until death or for a minimum of 90 days after the last patient was recruited.

population
Patients were recruited from seven centres across London, UK (St George's Hospital NHS Trust, Royal Marsden NHS Foundation Trust, Trinity Hospice, St Christopher's Hospice, St Helier Hospital, Princess Alice Hospice and the Sam Beare Unit). In order to be eligible for inclusion in this study, patients had to be at least 18 years old, be able to give written informed consent, have locally advanced or metastatic cancer, no longer be undergoing disease-modifying treatments and be under the care of a specialist palliative care team. Patients were excluded from the study if they were receiving or had received vitamin B12 injections within the previous 2 years or if they had known disease of the terminal ileum or pernicious anaemia.

procedures
After providing written informed consent, patients who agreed to participate in the study underwent the following assessments:

  • Donation of a 10-ml venous specimen of blood for analysis of serum vitamin B12 and CRP. If patients had had these assays taken within 24 h of entry into the study as part of their routine clinical care, then these tests were not repeated in order to avoid subjecting vulnerable participants to unnecessary investigations. Specimens were analysed at local clinical laboratories.
  • Eastern Cooperative Oncology Group performance status
  • Age
  • Gender
  • Disease- and treatment-related variables (primary diagnosis, number and sites of metastases, presence of clinically apparent infection, concurrent use of antibiotics and presence of comorbid diseases)
  • Date of death (obtained from scrutiny of medical records and contacting family doctors)

statistical methods
This study was designed and powered to be a confirmatory study of the prognostic accuracy of the BCI. It was determined that the BCI would be considered to be accurate as a prognostic index if the 95% confidence intervals (CIs) for 90-day survival obtained in this study contained the point estimates of survival reported in the original exploratory study [11]. Because the CIs in Geissbuhler et al. [11] study were relatively wide, a conventional power calculation was not undertaken. Instead, an estimation approach was adopted, i.e. the study set out to estimate the 90-day survival with 95% CI ± 10% in each prognostic group. Approximately 100 people are required in order to be able to estimate a 95% CI to within ± 10% (assuming a 50% survival rate). As the BCI prognostic categories were originally designed to split samples into three similarly sized groups, it was estimated that it would be necessary to recruit 300 patients in total.

The survival distributions of the three prognostic groups have been compared with pairwise log-rank {chi}2 tests. The survival distributions are graphically presented using Kaplan–Meier curves. Length of survival has been described using the median and 95% CIs. The proportion of patients who died on or before 90 days is also presented with 95% CIs. All statistical analysis was carried out using SPSS v13 for Windows.


   results
Top
 Abstract
 introduction
 patients and methods
 results
discussion
 Acknowledgements
 References
 
A total of 379 patients consented to participate in this study. However, in 50 patients the relevant blood tests were not taken or the results of the tests were incomplete. Thus, only 329 patients were included in the final analysis. Demographic and disease-related variables for these patients are shown in Table 1. The median age of participants was 68.7 years (range 23–94). No survival data were available for nine patients (data for these subjects was censored at the last known date that they were alive). The median survival for all subjects was 42 days (95% CI 34.8–49.2).


View this table:
[in this window]
[in a new window]

  		Table 1 Summary of patient characteristics

 
performance of the BCI
The subjects were divided into following three groups according to the BCI scores: 90/329 patients (27.3%) had a BCI score in the best prognostic category (BCI ≤10 000; group 1), 111/329 patients (33.7%) had a BCI score in the intermediate range (BCI 10 001–40 000; group 2) and 128/329 patients (38.9%) had a BCI score in the worst prognostic category (BCI >40 000; group 3). Kaplan–Meier survival curves for each of the prognostic categories are shown in Figure 1. The log-rank test for trend was statistically significant ({chi}2 = 18.38, P < 0.001).


Figure 1
View larger version (16K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1. Kaplan–Meier survival curves for vitamin B12/C-reactive protein Index groups.

 
Patients in group 3 had a significantly worse survival than group 1 (log-rank {chi}2 = 17.9, P < 0.001) and group 2 (log-rank {chi}2 = 7.0, P = 0.008). However, patients in group 1 did not have a significantly better prognosis than those in group 2 (log-rank {chi}2 = 2.9, P = 0.091). The median survival of patients in group 1 was 71 days (95% CI 45.9–96.2), of patients in group 2 was 43 days (95% CI 28.8–57.2 days) and of patients in group 3 was only 29 days (95% CI 22.0–36.1 days).

A comparison between the median 90-day mortality of patients in this study and the figures in Geissbuhler et al. [11] original study is shown in Table 2. The point estimates for 90-day mortality obtained from Geissbuhler et al. [11] study for each of the three BCI prognostic groups did not fall within the 95% CIs for the estimates obtained in our own study.


View this table:
[in this window]
[in a new window]

  		Table 2 Comparison between current study and Geissbuhler et al. [11] 2000

 

   discussion
Top
 Abstract
 introduction
 patients and methods
 results
 discussion
Acknowledgements
 References
 
The purpose of this confirmatory study was to determine the effectiveness of the BCI in an independent cohort of patients with advanced cancer. Our prior hypothesis had been that the 95% CIs for the 90-day survival of the three prognostic groups in our own study would include the point estimates for 90-day survival in the Geissbuhler study. However, we found that the mortality in our best prognostic group was higher than in Geissbuhler's study (58.9% versus 47.2%), whereas the mortality in the other two groups was lower than in the original study (64.0% versus 72.5% for group II and 78.9% versus 90.6% for Group III). These differences may partly be explained by differences in the populations studied. Although both studies involved palliative care patients, there is no universal agreement about what constitutes a palliative care population. Geissbuhler's study was concerned with consecutive cancer admissions to a geriatric unit. Our own study involved cancer patients under the care of specialist palliative care teams in general hospitals, cancer centres and hospices. Although the median survival time for the patients in both populations was broadly similar (45 days; 95% CI 32–56 and 42 days; 95% CI 34.8–49.2, respectively), there are reasons to suppose that the two populations may have differed in important characteristics. In Geissbuhler's study, blood tests were taken routinely from all admissions to the unit and it does not appear that specific consent was obtained from individuals before entry into the study (although ethics committee approval was obtained). Our own study only included patients who were able to give written informed consent. As a result, our population was likely to include fewer patients with an extremely short prognosis (delirium is an independent poor prognostic factor in palliative care patients [9]). It is also not clear whether any patients underwent disease-modifying treatments in Geissbuhler's original study, whereas in our own study all patients in whom further disease-modifying treatments were planned were excluded. Taken together, these two differences may explain why the range of 90-day mortality figures was broader in Geissbuhler's study (43.4%) than in our own (20%).

Nonetheless, our study provides at least partial support for the validity of the BCI as a prognostic index in patients with advanced cancer. Patients with a BCI of >40 000 had a significantly worse prognosis than patients with intermediate or low BCI values. However, the differences between the ‘good’ and the ‘intermediate’ prognostic groups failed to reach statistical significance. Thus, while a high BCI score had poor prognostic significance, an intermediate or low score was more difficult to interpret. A common management dilemma for palliative care specialists is determining which patients (who are unable to manage at home) would be more suitable for a terminal care admission to a hospice and which patients would be better placed in a nursing home. Although this decision will be based on a combination of factors (including symptom burden, social circumstances and the need for specialist interventions or nursing care), one important consideration is the expected prognosis. In general, hospices would prefer to admit patients with a prognosis of ‘weeks’ and would aim to transfer patients with longer prognoses to nursing or residential homes. Our study indicates that the BCI may help clinicians to make decisions about which patients have the poorest prognoses. However, the BCI alone is insufficiently accurate to be used in isolation and would need to be used in conjunction with clinician estimates and consideration of other prognostic factors [9, 12].

Only two published scales have previously undergone the recommended validation process of development in a training set of patients and confirmatory investigation in an independent testing set [9, 10]. The BCI can now be added to the short list of indices which have undergone confirmatory evaluation. The most widely used and best validated of these instruments is the Palliative Prognostic (PaP) score [13]. Pirovano and co-workers [13, 14] originally derived the scale from an exploratory study involving 519 palliative care patients. On multiple regression analysis, six variables (clinician prediction of survival, Karnofsky performance status, anorexia, dyspnoea, total white blood count and lymphocyte percentage) were found to be independently predictive of survival. The PaP score has subsequently been evaluated in independent patient populations [14, 15] and has been found to reliably divide patient populations into three prognostic groups, although one study has indicated that it is less good at distinguishing between patients with good or intermediate prognosis [15]. The other prognostic scale which has undergone appropriate validation is the Palliative Prognostic Index (PPI). This scale was developed in a ‘training’ set of 150 patients admitted to a hospice [16] and was subsequently evaluated in a ‘testing’ set consisting of 95 patients admitted to the same unit. The PPI consists of a weighted scale assessing five clinical variables (performance status, oral intake, oedema, dyspnoea at rest and delirium). Unlike the PaP, the PPI does not specifically include clinician estimates as an independent prognostic factor. However, a subsequent study [17] has reported that when clinicians are made aware of patients' PPI score, their own prognostic accuracy is significantly improved. In contrast to the PaP and the PPI, the attraction of the BCI is that it does not rely on subjective estimates of symptom severity or clinician estimates of survival.

There are biologically plausible mechanisms for understanding why raised CRP levels should be associated with a poorer prognosis [18]. CRP is a typical ‘acute phase protein’ and is raised in many chronic inflammatory conditions, cancer and infection. CRP is synthesised in hepatocytes in response to pro-inflammatory cytokines such as interleukin 6, interleukin 1 and tumour necrosis factor-{alpha}. Raised CRP levels have been reported to be a significant predictor of survival in many different tumour types [19–24]. Why raised serum vitamin B12 levels should be associated with a poorer prognosis is more difficult to explain. Elevated vitamin B12 levels are known to occur in patients with myeloproliferative disorders, hepatitis, cirrhosis, hepatocellular carcinoma and metastatic liver disease [25]. Carmel and Eisenberg [26] previously noted a relationship between vitamin B12 levels and survival in a heterogeneous (non-haematological) group of 139 cancer patients. The 24 patients with raised serum vitamin B12 levels had a median survival of 1 month compared with a median survival of 4 months for other subjects. There was a higher prevalence of metastatic disease in general and hepatic metastases in particular in the group with elevated vitamin B12 levels. It may be that a raised serum vitamin B12 level has prognostic significance because it is a reflection of disease extent. Geissbuhler et al. [11] postulated that elevated vitamin B12 levels may be secondary to elevated circulating transcobalamins (which can occur in many neoplastic or inflammatory conditions) combined with liver disease (either pre-existing disease or secondary to known or occult metastases).

There were a number of limitations to our study. Due to financial and practical constraints it was not possible to recruit a consecutive series of patients. No funding was specifically earmarked to undertake this project and so it was not feasible to employ dedicated personnel to screen every referral at each of the participating centres. In some patients (n = 50) who consented to participate in this study, data collection was incomplete. Reasons for failure to obtain data included patients who deteriorated or were discharged before the blood sample having been taken or results having been lost. However, the purpose of this study was to validate the BCI in a ‘real world’ environment. This emphasis on keeping the study simple, quick and practical also meant that blood samples were analysed locally in the routine clinical laboratories rather than being batched and analysed centrally. Although this means that there may have been some variation in the precision with which CRP and vitamin B12 levels were measured, we were keen to investigate whether the BCI was a practical instrument to use as part of routine clinical care, not only in a research setting. In retrospect, it would have been useful to include a clinician estimate of survival as part of the study assessments. This should be included in any future studies so that the ‘added-value’ of the BCI can be determined. Recent literature reviews have identified numerous putative individual prognostic factors in patients with advanced cancer [9, 12]. Further exploratory studies are warranted, that include all of the known prognostic factors, to develop new and better prognostic scales. In the light of this study, future research should include CRP and serum vitamin B12 assessments as part of the development process for any new prognostic scales.


   Acknowledgements
Top
 Abstract
 introduction
 patients and methods
 results
 discussion
 Acknowledgements
References
 
We would like to thank Rob Murphy, Eileen Rix, Alison Kennett, Karen Cook, Katherine Goller, Elizabeth Rees and Bridget Gwilliam for helping to collect data for this project. Thanks are also due to Janet Hardy, Julia Riley, Catherine Waight, Caroline Lucas, Martine Meyer, Victor Pace and Bernadette Lee for recruiting patients at the Royal Marsden Hospital (London and Surrey), Trinity Hospice, Princess Alice Hospice, St Helier Hospital, St Christopher's Hospice and the Sam Beare Unit, respectively. These units were part of the South London Research Group in Palliative Care.

Received for publication December 15, 2006. Revision received March 16, 2007. Accepted for publication March 23, 2007.


   References
Top
 Abstract
 introduction
 patients and methods
 results
 discussion
 Acknowledgements
 References
 
1. Christakis NA, Lamont EB. Extent and determinants of error in doctors' prognoses in terminally ill patients: prospective cohort study. Br Med J (2000) 320:469–472.[Abstract/Free Full Text]

2. Glare P, Virik K, Jones M, et al. A systematic review of physicians' survival predictions in terminally ill cancer patients. Br Med J (2003) 327:195.[Abstract/Free Full Text]

3. Weeks JC, Cook EF, O'Day SJ, et al. Relationship between cancer patients' predictions of prognosis and their treatment preferences [erratum appears in JAMA 2000 Jan 12; 283 (2): 203]. JAMA (1998) 279:1709–1714.[Abstract/Free Full Text]

4. Degner LF, Kristjanson LJ, Bowman D, et al. Information needs and decisional preferences in women with breast cancer. JAMA (1997) 277:1485–1492.[Abstract/Free Full Text]

5. Blanchard CG, Labrecque MS, Ruckdeschel JC, Blanchard EB. Information and decision-making preferences of hospitalized adult cancer patients. Soc Sci Med (1988) 27:1139–1145.[CrossRef][Web of Science][Medline]

6. den Daas N. Estimating length of survival in end-stage cancer: a review of the literature. J Pain Symptom Manage (1995) 10:548–555.[CrossRef][Web of Science][Medline]

7. The AM, Hak T, Koeter G, van Der Wal G. Collusion in doctor-patient communication about imminent death: an ethnographic study [comment]. Br Med J (2000) 321:1376–1381.[Abstract/Free Full Text]

8. Forster LE, Lynn J. Predicting life span for applicants to inpatient hospice. Arch Intern Med (1988) 148:2540–2543.[Abstract/Free Full Text]

9. Maltoni M, Caraceni A, Brunelli C, et al. Prognostic factors in advanced cancer patients: evidence-based clinical recommendations—a study by the Steering Committee of the European Association for Palliative Care. J Clin Oncol (2005) 23:6240–6248.[Abstract/Free Full Text]

10. Simon R, Altman DG. Statistical aspects of prognostic factor studies in oncology. Br J Cancer (1994) 69:979–985.[Web of Science][Medline]

11. Geissbuhler P, Mermillod B, Rapin CH. Elevated serum vitamin B12 levels associated with CRP as a predictive factor of mortality in palliative care cancer patients: a prospective study over five years. J Pain Symptom Manage (2000) 20:93–103.[CrossRef][Web of Science][Medline]

12. Vigano A, Dorgan M, Buckingham J, et al. Survival prediction in terminal cancer patients: a systematic review of the medical literature. Palliat Med (2000) 14:363–374.[Abstract/Free Full Text]

13. Pirovano M, Maltoni M, Nanni O, et al. A new palliative prognostic score: a first step for the staging of terminally ill cancer patients. Italian Multicenter and Study Group on Palliative Care. J Pain Symptom Manage (1999) 17:231–239.[CrossRef][Web of Science][Medline]

14. Maltoni M, Nanni O, Pirovano M, et al. Successful validation of the palliative prognostic score in terminally ill cancer patients. Italian Multicenter Study Group on Palliative Care. J Pain Symptom Manage (1999) 17:240–247.[CrossRef][Web of Science][Medline]

15. Glare P, Virik K. Independent prospective validation of the PaP score in terminally ill patients referred to a hospital-based palliative medicine consultation service. J Pain Symptom Manage (2001) 22:891–898.[CrossRef][Web of Science][Medline]

16. Morita T, Tsunoda J, Inoue S, Chihara S. The Palliative Prognostic Index: a scoring system for survival prediction of terminally ill cancer patients. Support Care Cancer (1999) 7:128–133.[CrossRef][Web of Science][Medline]

17. Morita T, Tsunoda J, Inoue S, Chihara S. Improved accuracy of physicians' survival prediction for terminally ill cancer patients using the Palliative Prognostic Index. Palliat Med (2001) 15:419–424.[Abstract/Free Full Text]

18. Mahmoud FA, Rivera NI. The role of C-reactive protein as a prognostic indicator in advanced cancer. Curr Oncol Rep (2002) 4:250–255.[Medline]

19. Pelliniemi TT, Irjala K, Mattila K, et al. Immunoreactive interleukin-6 and acute phase proteins as prognostic factors in multiple myeloma. Finnish Leukemia Group. Blood (1995) 85:765–771.[Abstract/Free Full Text]

20. Tartour E, Dorval T, Mosseri V, et al. Serum interleukin 6 and C-reactive protein levels correlate with resistance to IL-2 therapy and poor survival in melanoma patients. Br J Cancer (1994) 69:911–913.[Web of Science][Medline]

21. Legouffe E, Rodriguez C, Picot MC, et al. C-reactive protein serum level is a valuable and simple prognostic marker in non Hodgkin's lymphoma. Leuk Lymphoma (1998) 31:351–357.[Web of Science][Medline]

22. Kodama J, Miyagi Y, Seki N, et al. Serum C-reactive protein as a prognostic factor in patients with epithelial ovarian cancer. Eur J Obstet Gynecol Reprod Biol (1999) 82:107–110.[CrossRef][Web of Science][Medline]

23. Ljungberg B, Grankvist K, Rasmuson T. Serum acute phase reactants and prognosis in renal cell carcinoma. Cancer (1995) 76:1435–1439.[CrossRef][Web of Science][Medline]

24. O'Gorman P, McMillan DC, McArdle CS. Prognostic factors in advanced gastrointestinal cancer patients with weight loss. Nutr Cancer (2000) 37:36–40.[CrossRef][Web of Science][Medline]

25. Ermens AAM, Vlasveld LT, Lindemans J. Significance of elevated cobalamin (vitamin B12) levels in blood. Clin Biochem (2003) 36:585–590.[CrossRef][Web of Science][Medline]

26. Carmel R, Eisenberg L. Serum vitamin B12 and transcobalamin abnormalities in patients with cancer. Cancer (1977) 40:1348–1353.[CrossRef][Web of Science][Medline]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Palliat MedHome page
P Stone, L Kelly, R Head, and S White
Development and validation of a prognostic scale for use in patients with advanced cancer
Palliative Medicine, September 1, 2008; 22(6): 711 - 717.
[Abstract] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
18/8/1395    most recent
mdm138v1
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Kelly, L
Right arrow Articles by Stone, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kelly, L
Right arrow Articles by Stone, P.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?