Skip Navigation


NDT Advance Access originally published online on September 23, 2006
Nephrology Dialysis Transplantation 2007 22(1):187-195; doi:10.1093/ndt/gfl502
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
22/1/187    most recent
gfl502v1
Right arrow Alert me when this article is cited
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 ISI Web of Science
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 arrow Search for citing articles in:
ISI Web of Science (10)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by van Manen, J. G.
Right arrow Articles by Jager, K. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by van Manen, J. G.
Right arrow Articles by Jager, K. J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author [2006]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Confounding effect of comorbidity in survival studies in patients on renal replacement therapy

Jeannette G. van Manen1, Paul C. W. van Dijk2, Vianda S. Stel2, Friedo W. Dekker1, Montse Clèries3, Ferruccio Conte4, Terry Feest5, Reinhard Kramar6, Torbjørn Leivestad7, J. Douglas Briggs2, Bénédicte Stengel8 and Kitty J. Jager2

1Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, 2ERA–EDTA Registry, Dept. of Medical Informatics, Academic Medical Center – University of Amsterdam, Amsterdam, The Netherlands, 3Catalan Renal Registry (RMRC), Catalan Transplant Organisation, Catalan Health Service, Autonomous Government of Catalonia, Barcelona, Spain, 4Registro Lombardo Dialisi e Trapianto, "Ospedale UBOLDO", Cernusco S/N, Milano, Italy, 5The UK Renal Registry, Southmead Hospital, Bristol, UK, 6Austrian Dialysis and Transplant Registry (OEDTR), General Hospital of Wels, Wels, Austria, 7Norwegian Renal Registry, Institute of Immunology, Rikshospitalet University Hospital, Oslo, Norway and 8Institut National de la Santé et de la Recherche Médicale, INSERM Unité 258, Villejuif, France

Correspondence and offprint requests to: Kitty J. Jager, ERA-EDTA Registry, Department of Medical Informatics, J1B-125, Academic Medical Center, PO Box 22700, 1100 DE Amsterdam, The Netherlands. Email: k.j.jager{at}amc.uva.nl



   Abstract
Top
 Abstract
Introduction
 Subjects and methods
 Results
 Discussion
 Appendix 1
 Acknowledgements
 References
 
Background. After taking other confounding factors into account, the impact of comorbidity on mortality was investigated when comparing mortality between five European countries, dialysis modalities and renal disease groups.

Methods. The study included 15 571 incident patients on renal replacement therapy (RRT) from five national or regional registries participating in the European Renal Association—European Dialysis and Transplant Association Registry that collect comorbidity data. The presence of diabetes mellitus, ischaemic heart disease, peripheral vascular disease, cerebrovascular disease and malignancy was recorded at the start of RRT.

Results. The comorbidities were each independently associated with mortality, with hazard ratios (HRs) ranging from 1.40 (95% CI: 1.30–1.51) for peripheral vascular disease to 1.65 (95% CI: 1.48–1.83) for diabetes. Age, gender, primary renal disease, modality and country together explained 14.4% of the variance in mortality; the comorbidities explained an additional 1.9%. In the comparison of renal vascular disease with glomerulonephritis, the crude HR of 2.40 (95% CI: 2.12–2.72) changed to 1.24 (95% CI: 1.09–1.41) after adjustment for age, gender, primary renal disease, treatment modality and country and to 1.06 (95% CI: 0.93–1.22) after further adjustment for the comorbidities. For the comparison between countries and other patient groups, the change in the survival estimate after adjustment for comorbidity was less.

Conclusion. Comorbidity is an important predictor for mortality. However, after adjustment for age, gender, primary renal disease, treatment modality and country, when comparing outcomes between patient groups the influence of comorbidity may be less important than expected.

Keywords: comorbidity; confounding; renal replacement therapy; survival



   Introduction
Top
 Abstract
 Introduction
Subjects and methods
 Results
 Discussion
 Appendix 1
 Acknowledgements
 References
 
In clinical research among patients on renal replacement therapy, it is often investigated whether certain variables have an effect on mortality. In order to estimate the true effect of a single variable—like country, dialysis modality or primary renal disease—it is essential that the results are adjusted for the potential confounding effect of other variables—such as age—that are known to be highly predictive of mortality. If groups differ with respect to age, this may hamper a fair comparison of mortality between those groups. Adjustment for confounding by age will remove this effect.

Most of the time, group comparison will be adjusted for (possible) differences in age. Depending on the variable under study, adjustment for other potential confounders—like gender or primary renal disease—on top of age may be required. For many comparisons, adjustment for comorbidity is thought to be essential, since comorbidity has been shown to be predictive of mortality [1–11]. Consequently, many papers have stressed the need to collect data on comorbidity. However, collecting data on comorbidity is relatively difficult and time-consuming when comparing patient characteristics such as age or primary renal disease. As age and comorbidity are related, it may be that adjustment for comorbidity on top of age—and other potential confounders—may have limited value. The question was therefore raised, whether adjustment for comorbidity is always necessary.

Although many studies have reported on the association between comorbidity and mortality, very few of them have quantified the actual bias resulting from lack of adjustment for comorbidity in survival comparisons. More importantly, the bias resulting from lack of adjustment for comorbidity, on top of other characteristics, has not been the object of study so far.

To our knowledge, the Dialysis Outcomes and Practice Patterns Study (DOPPS) is the only study that has evaluated how much the observed differences in mortality between dialysis centres and countries can be attributed to case-mix factors including comorbidity [12]. It was found that the variability in demographic factors and comorbid conditions explained a rather small part of the differences in mortality, indicating that the importance of adjustment for these factors may be limited in practice. Within DOPPS the differences in case-mix between the countries may not have been large enough to induce substantial confounding. Possibly, the adjustment for case-mix, therefore, resulted in only a limited change of relative risks of deaths. In addition, the separate contribution of comorbidity to these results was not presented; e.g. it may be that it is far more important to adjust for comorbidity than for gender.

In the present study, our aim was to quantify the confounding effect of comorbidity when studying differences in survival between countries, dialysis modalities or renal disease groups. The study was performed in incident patients on renal replacement therapy for end-stage renal disease participating in the European Renal Association–European Dialysis and Transplant Association (ERA–EDTA) Registry.



   Subjects and methods
Top
 Abstract
 Introduction
 Subjects and methods
Results
 Discussion
 Appendix 1
 Acknowledgements
 References
 
Subjects
Data were collected from the five national and regional renal registries participating in the ERA-EDTA Registry that were able to provide data on comorbidity for the purpose of the present study. Comorbidity is not routinely registered in many countries participating in the ERA-EDTA Registry. Participating registries were as follows: Austrian Dialysis and Transplant Registry (OEDTR) (1998–2001), Catalan Renal Registry (Registre de Malalts Renals de Catalunya—RMRC) (Catalonia, Spain) (1995–2001), Lombardy Registry of Dialysis and Transplantation (Registro Lombardo Dialisi e Trapianto) (Lombardy, Italy) (1994–1997), Norwegian Renal Registry (1999–2001), Renal Association UK Renal Registry (United Kingdom, England/Wales) (selected sample of the database 1999–2001). Numbers in brackets indicate the time period of the incident cohort used. Incident patients >18 years of age at the start of renal replacement therapy (RRT) and surviving the first 91 days since start of RRT were included. For inclusion of patient data into the national and regional registries, these organizations complied with their national data protection legislation.

Classification of comorbid conditions
The definitions used by each participating registry for diabetes mellitus (both as a renal disease and as a comorbidity in addition to another renal disease), ischaemic heart disease, peripheral vascular disease, cerebro-vascular disease and malignancy are shown in Appendix 1. Data collection on comorbidity was performed by a doctor or high-skilled nurse. Although detailed descriptions of diagnostic criteria for each comorbidity were often lacking, in all the cases the comorbid conditions were indicated as being present or absent in the medical history at the start of RRT. Sometimes they were further specified by providing subcategories of the disease involved. In all registries it was possible to distinguish between ‘no comorbidity’ or ‘missing data on comorbidity’, with the exception of Norway. In Norway, all patients without any data were considered to have no comorbidity.

Data analyses
To examine the association between the presence of comorbid conditions and survival time, Cox proportional hazards regression analysis was used. A Cox regression analysis provides a hazard ratio (HR) that indicates the instantaneous chance of death for a risk group compared with a reference group. For example, an HR of 2.0 for men compared with women means that men have a risk of death that is twice as high as that of women. If men and women differ with respect to age in a study population, a Cox regression analysis would provide the opportunity to adjust for the confounding effect of age. Instead of using only gender in such a Cox model, one would also include the confounder age in order to yield an HR for gender that is adjusted for age, i.e. indicating the risk of death for men vs women irrespective of their age.

The predictive values of comorbidity and other patient characteristics were expressed as the percentage of explained variance in survival (R2) and was calculated according to Nagelkerke [13]. R2 expresses the extent to which the variables included in a Cox regression model explain the variance in survival time. The higher the percentage the better, with 0% indicating no prediction at all and 100% indicating perfect prediction. In this study, the effect of additional adjustment for comorbidity was investigated by studying (i) the increase in R2 for Cox regression models including an increasing number of (demographic or clinical) variables, and finally, including comorbidity, and (ii) the HRs of Cox models investigating the relationship of country, treatment modality and renal disease on the one hand with mortality on the other hand. For the latter purpose, the crude as well as the adjusted HRs were calculated and three models were built: (i) the crude model, only including the one variable of interest—i.e. country, dialysis modality or primary renal disease, (ii) the adjusted model, including the variable of interest and the relevant potential confounders—i.e. age, gender, primary renal disease, treatment modality and country (the relevant potential confounders that were adjusted for did of course not include the variable of interest in a particular model, see Table 1), (iii) the adjusted model, including the variable of interest, the relevant potential confounders and the comorbidities.


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

  		Table 1. Effect of adjustment for comorbidity on survival effect estimates on top of other relevant potential confounders. Comparison by predictor variables—i.e. country, dialysis modality and two primary renal diseases

 
In the crude analysis of the relationship between diabetes and mortality (Table 2), the comorbidity variable ‘diabetes’ also included diabetes as the cause of renal failure. In the multivariate analyses (concerning the adjusted models), ‘diabetes’ as a comorbidity did not include diabetes as the cause of renal failure, since diabetes as the cause of renal failure was already included in the variable ‘primary renal disease’. Thus it was investigated to what extent comorbidity adds to adjustment for primary renal disease and other characteristics


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

  		Table 2. Influence of each comorbidity on survival

 
In the Cox regression analysis, death was the event of interest and follow-up time was censored at recovery of renal function, loss of follow-up, at the end of the observation period [31 December 2001 for Austria, Norway, Spain (Catalonia) and United Kingdom (England/Wales); 31 December 1997 for Italy (Lombardy)], and when patients were still alive and on RRT after 4 years.

There was no uniform follow-up period for the countries because the follow-up time that was registered was different for each country. In order to avoid reduction of informative data it was decided not to censor follow-up time after 2 years (available follow-up time for each country); as the proportional hazard assumption was checked and appeared to be valid for the data of our study, this does not invalidate the analysis.

All analyses were performed using SPSS 11.0.1.



   Results
Top
 Abstract
 Introduction
 Subjects and methods
 Results
Discussion
 Appendix 1
 Acknowledgements
 References
 
In the study, 15 571 patients from five renal registries were included (Table 3). The mean age of the patients was 61.6 years, 61.5% were males and 20.0% had diabetes as primary renal disease. Of the study patients, 78.8% were on haemodialysis (HD), 18.8% on peritoneal dialysis (PD) and 2.5% were living on a functioning graft.


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

  		Table 3. General characteristics of the study population

 
The most common comorbidity was diabetes (27.0%); in 7.0% diabetes was not the cause of renal failure. Malignancies were the least common (10.1%). The prevalence of each comorbidity strongly increased with increasing age. For example, the prevalence of ischaemic heart disease was 2.4% for patients aged 18–40 years, 14.1% in patients aged 40–60 years and 29.0% in patients over 60. The prevalence of comorbidities differed somewhat between countries, especially with regard to diabetes, which varied from 20.2% in Lombardy (Italy) to 39.8% in Austria, and with regard to peripheral vascular disease, which varied from 13.3% in England/Wales (UK) to 34.1% in Austria. As expected, the prevalence of heart disease and peripheral disease was higher in countries with a high prevalence of diabetes mellitus.

In Table 2 the association between each comorbidity and mortality is shown. The crude HRs ranged from 1.99 (95% CI: 1.86–2.12) for diabetes to 2.24 (95% CI: 2.05–2.44) for malignancies. The HRs when adjusted for age, gender, primary kidney disease, modality and country were lower and ranged from 1.40 (95% CI: 1.30–1.51) for peripheral vascular disease to 1.65 (95% CI: 1.48–1.83) for diabetes.

It was investigated to what extent the comorbidities were predictive of mortality after adjustment for other general characteristics. It appeared that age alone explained 11.2% of the variance (Figure 1). Subsequent addition of gender, primary renal disease, treatment modality and country to the model showed that these explained 0, 2.7, 0.2 and 0.3% in addition to age, respectively, resulted in a total explained variance of 14.4%. On top of these factors, the five comorbidities explained an additional 1.9% variance, resulting in 16.3% explained variance by all factors. The same patterns were found in each country (data not shown). Also, for the three age classes the patterns were rather similar, except that within each age class the percentage explained variance by age was much smaller.


Figure 1
View larger version (18K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Prediction of survival by general characteristics and comorbidity. n = 15376 (195 missing cases due to one or more missing variables). Age, gender, primary renal disease, treatment modality, country and five comorbidities were consecutively entered into the model. The amount of variance explained by a factor reflects additional explained variance on top of the factors already included in the model. aprd, primary renal disease.

 
In Figure 1, the percentage of explained variance is given for the five comorbidities together. The total amount of explained variance by all comorbidities is relatively small and there is no consistency in the importance of comorbidities across countries and age groups. Thus there is no apparent hierarchy in the comorbidities.

Finally, it was investigated to what extent adjustment for comorbidity in survival analyses actually influences HRs for three different predictor variables of interest (country, dialysis modality and primary renal disease) after correcting for other differences in case-mix. The differences in case-mix between the countries can be seen in Table 3. The differences in case-mix between the dialysis modalities HD and PD were mean age, 63 and 60; percentage males, 62 and 61; percentage diabetes, 20 and 19; percentage with at least one comorbidity, 50 and 42. The differences in case-mix between the primary renal diseases, renal vascular disease and glomerulonephritis were mean age, 68 and 55; percentage males, 70 and 70; percentage HD, 81 and 77; percentage with at least one comorbidity, 65 and 30. Thus, the differences in case-mix were relatively large between patients with glomerulonephritis and patients with renal vascular disease, especially with respect to age and comorbidity.

The HRs of the different classes of the predictor variables of interest were compared (Table 1). The crude HRs of the countries ranged from 0.99 (95% CI: 0.86–1.14) for Lombardy (Italy) to 1.49 (95% CI: 1.30–1.71) for Austria, when compared with the UK (England/Wales). Adjusting the HRs for age, gender, primary renal disease and treatment modality reduced the differences between countries, now ranging from 0.97 (95% CI: 0.84–1.12) for Lombardy (Italy) to 1.39 (95% CI: 1.15–1.68) for Norway. The differences between the HRs were only slightly further reduced when the analysis was additionally adjusted for the five comorbidities: the HRs ranged from 0.99 (95% CI: 0.86–1.14) for Lombardy (Italy) to 1.37 (1.14–1.66) for Norway.

The same adjustments were made for the comparison of the two major treatment modalities: HD and PD. The crude HR for HD compared with PD was 1.18 (95% CI: 1.08–1.29). This HR decreased to 0.99 (95% CI: 0.90–1.08) after adjustment for general characteristics. Additional adjustment for comorbidity revealed an HR of 0.96 (95% CI: 0.88–1.06).

This procedure was also followed for the comparison of mortality between patients with two different causes of renal failure with a large difference in prognosis, i.e. glomerulonephritis and renal vascular disease. The crude HR was 2.40 (95% CI: 2.12–2.72) and decreased to 1.24 (95% CI: 1.09–1.41) after adjustment for general characteristics and further decreased to 1.06 (95% CI: 0.93–1.22) after further adjustment for comorbidity.



   Discussion
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
Appendix 1
 Acknowledgements
 References
 
This study among 15 571 patients on renal replacement therapy from five European countries confirms that comorbidity is an important predictor for survival and is thus of prognostic importance for the individual patient. The data of the present study also show that in the comparison of survival between five European countries, two dialysis modalities, and two primary renal disease groups, the estimate of the HRs changed importantly after adjustment for general characteristics like age or gender, but very little after further adjustment for comorbidity.

Since the present study as well as many other studies have shown that comorbidity is clearly associated with mortality, it is initially surprising that adjustment for comorbidity—on top of other characteristics—has a limited effect on mortality comparisons. It may simply be that age and comorbidity are highly correlated—as is well known and was also demonstrated in this study. The same holds true for primary renal disease. Adjusting for age (or primary renal disease), therefore, already (partly) removes the effect of comorbidity. Additional adjustment for comorbidity on top of age or primary renal disease will then have a limited effect. Age is effectively a proxy for comorbidity. Additional analyses revealed that comorbidity alone indeed explained a larger part of the variance (6.9%) than it did on top of age and other characteristics (1.9%). Age alone explained 11.2% of the variance.

A second reason for the limited effect of additional adjustment for comorbidity may be the limited number of comorbidities available from all five registries, variation in definitions and absence of severity grading of the comorbidities. Adjustment for a larger number of comorbidities and for the severity of comorbidity could have had a larger effect on HR. To investigate these possibilities the data from the Necosad study were analysed.

The Necosad study is a multi-centre study on 1489 incident dialysis patients treated in the Netherlands for which 12 comorbidities and severity grading are recorded [5], of which seven were not recorded in the five European countries. These additional seven comorbidities explained an additional 2.6% of the variance in survival, on top of the 19% that was explained by general characteristics and the five comorbidities. However, additional adjustment for these seven comorbidities in the comparisons of survival between renal vascular disease and glomerulonephritis—the two groups that differed most with respect to comorbidity—only slightly changed the estimate of the HRs [HR changed from 2.18 (95% CI: 1.40–3.39) when adjusting for general characteristics and five comorbidities to 2.34 (95%CI: 1.49–3.67) when adjusting for general characteristics and 12 comorbidities]. Thus there was little additional benefit for the effort of collecting the additional seven comorbidities, and grading severity. These analyses on Necosad, a carefully conducted observational study, support the findings of this study.

It is well known that data based on the clinical diagnosis of a disease are subject to misclassification and lack sensitivity [14]. It may be that with better recording and grading of comorbidity there would be larger effects on the estimate of the HR, but in practical terms for registries and most clinical studies this is not practicable. This study shows that the effect of adjustment for comorbidity—in the way that it is recorded in actual practice—is relatively limited.

The comorbidities were not recorded completely uniformly in all countries participating in the study. Due to the resulting statistical noise, the real effect of adjustment for comorbidity may have been slightly underestimated. However, this study investigated the effect of adjustment for comorbidity in the way it is recorded in daily practice.

The maximum percentage of variance that could be explained by comorbidity and other covariates together was low and did not exceed 22%. This may be interpreted as a sign of low data quality. However, other studies have shown that the ability to predict survival time is relatively limited [15,16], even when the number of variables to adjust for is very large. This low percentage of explained variance seems inherent to the Cox regression analysis used in this and other studies. Experts on survival analysis state that values of about 40% can be reached, but that—depending on the methods of calculation—studies of survival often have much lower estimates of explained variation [17]. Moreover, the purpose of studies is not to accurately predict survival in individuals but to detect clinically meaningful differences in survival time of months or years between groups. Consequently, the low percentage of explained variance found in this study is probably not an indicator of limitations in data.

In theory it will be possible that adjustment for comorbidity has a larger effect on the survival estimates when comparing other groups than those presented in this study. For example, in this study population the prevalence of comorbidity was similar between groups within different age strata. If comorbidity differed more between groups within different age strata, additional adjustment for comorbidity might have had a stronger effect.

Since there was only one country with a longer follow-up period than 4 years, we censored the follow-up for the patients of this country at 4 years. Censoring earlier in follow-up, i.e. 2 years for all countries, did not change the results of the study. Consequently, we may assume that the results are not influenced by follow-up time.

This appears to be the first study showing the actual influence of adjustment for comorbidity on HR estimates on all patients starting RRT. Recently, DOPPS on 16 720 prevalent patients in the US, Europe and Japan selected for HD only showed that the majority of the 22 recorded comorbidities were associated with mortality, and that adjustment for these comorbidities together with other case-mix factors decreased differences in mortality between countries to some extent [12]. However, the additional value of comorbidity in the prediction of survival was not investigated separately from the effects of the other case-mix factors.

In conclusion, this study confirms that comorbidity is common in dialysis patients and that comorbidities are clearly associated with survival. However, after adjustment for age, gender, primary renal disease, treatment modality and country, when comparing outcomes between patient groups the influence of comorbidity, may be less important than expected. Thus studies that are not able to record comorbidity may already account for a large part of the confounding effect of comorbidity by adjusting for age and primary renal disease. Some residual confounding may be inevitable since it is difficult—and maybe impossible—to record comorbidity in a way that does justice to its complexity. Furthermore, if comorbidity differs between groups within different age strata, its confounding effect will be less adequately removed by adjustment for age, and in those cases additional adjustment for comorbidity might be required. Nevertheless, collection of comorbidity data where possible is desirable in order to obtain the most valid effect estimate possible in RRT survival studies, and for other reasons such as estimation of disease burden or defining patient groups for research purposes.



   Appendix 1
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 Appendix 1
Acknowledgements
 References
 
Classification of comorbid conditions used by renal registries

Classification of co-morbidity Definition or more detailed description of the classification of comorbidity
Diabetes mellitus
Austria Diabetes mellitus type I Insulin needed within year after diagnosis
Diabetes mellitus type II Age at start of diabetes mellitus >40 years, treatment with oral antidiabetics or diet possible for longer period of time
Secondary diabetes mellitus Caused by other conditions (pancreatic diseases, hormonal imbalance, drug or chemically induced, insulin receptor abnormality, genetic syndrome)
Unspecified diabetes mellitus Note: Transitory steroid-induced diabetes mellitus should not be classified as diabetes mellitus
Catalonia (Spain) Diabetes mellitus type I
Diabetes mellitus type II
Secondary or unspecified diabetes mellitus
England/Wales (UK) Diabetes mellitus Including diet controlled and drug induced diabetes mellitus
Lombardy (Italy) Diabetes mellitus type I
Diabetes mellitus type II
Secondary or unspecified diabetes mellitus
Norway Diabetes mellitus type I Insulin dependency from debut (not secondary)
Diabetes mellitus type II
Secondary diabetes mellitus
Unspecified diabetes mellitus
Ischaemic heart disease
Austria Coronary heart disease Documented by angiography, stress echo, thallium scintigraphy, etc.
Myocardial infarction
Angina pectoris
Catalonia (Spain)
ICD-9-CMa code 410-414
410:Acute myocardial infarction;
411:Other forms of subacute ischaemic heart disease;
412:Old myocardial infarction;
413:Angina pectoris;
414:Other forms of chronic ischaemic heart disease
England/Wales (UK) CABG or coronary angioplasty
Myocardial infarction Diagnosed by ST segment evaluation, Q waves in relevant leads, enzyme rise > 2x upper limit of normal (or rise in creatinine kinase-MB above local reference range)
Angina pectoris History of chest pain on exercise with or without ECG changes, exercise tolerance test, radionucleotide imaging or angiography
Lombardy (Italy) Coronary heart disease
Myocardial infarction
Angina pectoris
Norway Myocardial infarction
Angina pectoris
Peripheral vascular disease
Austria Peripheral vascular disease Included aortic aneurysm, diffuse vascular calcifications, documented vascular stenosis
Catalonia (Spain)
ICD-9-CMa code 440-441, 443
440:Atherosclerosis;
441:Aortic aneurysm and dissection;
443:Other peripheral vascular disease
England/Wales (UK) Claudication Current claudication based on a history, with or without Doppler or angiographic evidence
Angioplasty, stenting vascular graft (non coronary) Includes vascular grafts (e.g. aortic bifurcation grafts) and renal artery stents
Amputation for peripheral vascular disease
Lombardy (Italy) Peripheral vascular disease
Norway Peripheral vascular disease Including all arteries (except carotid/cerebrovascular and coronary)
Should include both symptomatic disease and affection necessitating pre-transplant intervention
Cerebrovascular disease
Austria Cerebrovascular disease
Catalonia (Spain)
ICD-9-CMa code 430-438, 342
430:Subarachnoid haemorrhage;
431:Intracerebral haemorrhage;
432:Other and unspecified intracranial haemorrhage;
433:Occlusion and stenosis of precerebral arteries;
434:Occlusion of cerebral arteries;
435:Transient cerebral ischaemia;
436:Acute but ill-defined cerebrovascular disease;
437:Other and ill-defined cerebrovascular disease;
438:Late effects of cerebrovascular disease;
342:Hemiplegia and hemiparesis
England/Wales (UK) Cerebrovascular disease Any history of strokes (whatever cause) and including transient ischaemic attacks caused by carotid disease
Lombardy (Italy) Cerebrovascular disease
Norway Cerebrovascular disease
Malignancy
Austria Malignancy Solid tumour and other malignancies, including basal cell carcinoma
Catalonia (Spain) ICD-9-CMa code 140-208, 230-239 Including basal cell carcinoma
England/Wales (UK) Malignancy Any history of malignancy (even if curative) e.g. removal of melanoma, excluding basal cell carcinoma
Lombardy (Italy) Malignancy Including basal cell carcinoma
Norway Malignancy Including basal cell carcinoma

aUnited States National Center for Health Statistics. Annotated International Classification of Diseases, Ninth Revision, Clinical Modification. Ann Arbor: Edwards Brothers, Inc., 1987.



   Acknowledgements
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 Appendix 1
 Acknowledgements
References
 
We would like to thank the patients and staff of all the dialysis and transplant units who have contributed data via their national and regional renal registries.

We also would like to thank Dr H. K. Stumvoll [Austrian Dialysis and Transplant Registry (OEDTR)], Dr E. Vela [Catalan Transplant Organization (OCATT)], Dr D. Ansell (the UK Renal Registry) and the ERA-EDTA Registry committee members and staff members for their advice in the analysis and drafting of the article.

The ERA–EDTA Registry is funded by ERA–EDTA. The following companies have committed funds in the form of unrestricted educational grants to assist the ERA–EDTA in the financial support of the Registry: Amgen, Baxter, Fresenius Medical Care, Gambro, Hoffmann-La Roche, Hospal, Ortho-Biotech and Shire.

Conflict of interest statement. None declared.



   References
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 Appendix 1
 Acknowledgements
 References
 

  1. Miskulin DC, Martin AA, Brown R, et al. (2004) Predicting 1 year mortality in an outpatient haemodialysis population: a comparison of comorbidity instruments. Nephrol Dial Transplant 19:413–420.[Abstract/Free Full Text]
  2. Beddhu S, Bruns FJ, Saul M, Seddon P, Zeidel ML. (2000) A simple comorbidity scale predicts clinical outcomes and costs in dialysis patients. Am J Med 108:609–613.[CrossRef][Web of Science][Medline]
  3. Davies SJ, Phillips L, Naish PF, Russell GI. (2002) Quantifying comorbidity in peritoneal dialysis patients and its relationship to other predictors of survival. Nephrol Dial Transplant 17:1085–92.[Abstract/Free Full Text]
  4. Khan IH, Catto GRD, Edward N, Fleming LW, Henderson IS, MacLeod AM. (1993) Influence of coexisting disease on survival on renal replacement therapy. Lancet 341:415–418.[CrossRef][Web of Science][Medline]
  5. van Manen JG, Korevaar JC, Dekker FW, Boeschoten EW, Bossuyt PMM, Krediet RT. (2002) How to adjust for comorbidity in survival studies in ESRD patients: a comparison of different indices. Am J Kidney Dis 40:82–89.[CrossRef][Web of Science][Medline]
  6. Fried L, Bernardini J, Piraino B. (2001) Charlson comorbidity index as a predictor of outcomes in incident peritoneal dialysis patients. Am J Kidney Dis 37:337–342.[Web of Science][Medline]
  7. Charlson ME, Pompei P, Ales KL, MacKenzie CR. (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40:373–383.[CrossRef][Web of Science][Medline]
  8. Hemmelgarn BR, Manns BJ, Quan H, Ghali WA. (2003) Adapting the charlson comorbidity index for use in patients with ESRD. Am J Kidney Dis 42:125–132.[CrossRef][Web of Science][Medline]
  9. van Manen JG, Korevaar JC, Dekker FW, Boeschoten EW, Bossuyt PM, Krediet RT. (2003) Adjustment for comorbidity in studies on health status in ESRD patients: which comorbidity index to use? J Am Soc Nephrol 14:478–485.[Abstract/Free Full Text]
  10. Miskulin DC, Meyer KB, Martin AA, et al. (2003) Comorbidity and its change predict survival in incident dialysis patients. Am J Kidney Dis 41:149–161.[CrossRef][Web of Science][Medline]
  11. Marcelli D, Stannard D, Conte F, Held PJ, Locatelli F, Port FK. (1996) ESRD patient mortality with adjustment for comorbid conditions in Lombardy (Italy) versus the United States. Kidney Int 50:1013–1018.[Web of Science][Medline]
  12. Goodkin DA, Bragg-Gresham JL, Koenig KG, et al. (2003) Association of Comorbid Conditions and Mortality in Hemodialysis Patients in Europe, Japan, and the United States: The Dialysis Outcomes and Practice Patterns Study (DOPPS). J Am Soc Nephrol 14:3270–3277.[Abstract/Free Full Text]
  13. Nagelkerke NJD. (1991) A note on a general definition of the coefficient of determination. Biometrika 78:691–692.[Abstract/Free Full Text]
  14. Longenecker JC, Coresh J, Klag MJ, et al. (2000) Validation of comorbid conditions on the end-stage renal disease medical evidence report: the CHOICE study. Choices for Healthy Outcomes in Caring for ESRD. J Am Soc Nephrol 11:520–529.[Abstract/Free Full Text]
  15. Mallamaci F, Tripepi G, Cutrupi S, Malatino LS, Zoccali C. (2005) Prognostic value of combined use of biomarkers of inflammation, endothelial dysfunction, and myocardiopathy in patients with ESRD. Kidney Int 67:2330–2337.[CrossRef][Web of Science][Medline]
  16. Jager KJ, Merkus MP, Dekker FW, et al. (1999) Mortality and technique failure in patients starting chronic peritoneal dialysis: results of the Netherlands Cooperative Study on the Adequacy of Dialysis. Kidney Int 55:1476–1485.[CrossRef][Web of Science][Medline]
  17. Schemper M and Stare J. (1996) Explained variation in survival analysis. Stat Med 15:1999–2012.[CrossRef][Web of Science][Medline]
Received for publication: 11. 1.06
Accepted in revised form: 26. 7.06


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
 Nephrol Dial TransplantHome page
S. Sawhney, O. Djurdjev, K. Simpson, A. Macleod, and A. Levin
Reply
Nephrol. Dial. Transplant., December 1, 2009; 24(12): 3896 - 3896.
[Full Text] [PDF]

Home page
 CJASNHome page
D. Miskulin, J. Bragg-Gresham, B. W. Gillespie, F. Tentori, R. L. Pisoni, H. Tighiouart, A. S. Levey, and F. K. Port
Key Comorbid Conditions that Are Predictive of Survival among Hemodialysis Patients
Clin. J. Am. Soc. Nephrol., November 1, 2009; 4(11): 1818 - 1826.
[Abstract] [Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
S. Sawhney, O. Djurdjev, K. Simpson, A. Macleod, and A. Levin
Survival and dialysis initiation: comparing British Columbia and Scotland registries
Nephrol. Dial. Transplant., October 1, 2009; 24(10): 3186 - 3192.
[Abstract] [Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
L. Karamadoukis, D. Ansell, R. N. Foley, S. P. McDonald, C. R. V. Tomson, L. Trpeski, and F. J. Caskey
Towards case-mix-adjusted international renal registry comparisons: how can we improve data collection practice?
Nephrol. Dial. Transplant., August 1, 2009; 24(8): 2306 - 2311.
[Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
K. J. Jager and C. Zoccali
Comorbidity data collection by renal registries--a remaining challenge
Nephrol. Dial. Transplant., August 1, 2009; 24(8): 2311 - 2313.
[Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
P. Saudan, M. Kossovsky, G. Halabi, P. Y. Martin, T. V. Perneger, and for the Western Switzerland Dialysis Study Group
Quality of care and survival of haemodialysed patients in western Switzerland
Nephrol. Dial. Transplant., June 1, 2008; 23(6): 1975 - 1981.
[Abstract] [Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
C. Couchoud, O. Moranne, L. Frimat, M. Labeeuw, V. Allot, and B. Stengel
Associations between comorbidities, treatment choice and outcome in the elderly with end-stage renal disease
Nephrol. Dial. Transplant., November 1, 2007; 22(11): 3246 - 3254.
[Abstract] [Full Text] [PDF]

Home page
 J. Am. Soc. Nephrol.Home page
F. Tentori, W. C. Hunt, M. Rohrscheib, M. Zhu, C. A. Stidley, K. Servilla, D. Miskulin, K. B. Meyer, E. J. Bedrick, H. K. Johnson, et al.
Which Targets in Clinical Practice Guidelines Are Associated with Improved Survival in a Large Dialysis Organization?
J. Am. Soc. Nephrol., August 1, 2007; 18(8): 2377 - 2384.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
22/1/187    most recent
gfl502v1
Right arrow Alert me when this article is cited
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 ISI Web of Science
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 arrow Search for citing articles in:
ISI Web of Science (10)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by van Manen, J. G.
Right arrow Articles by Jager, K. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by van Manen, J. G.
Right arrow Articles by Jager, K. J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?