Comorbidities and survival of patients with type 1 diabetes on renal replacement therapy.

AIMS/HYPOTHESIS: Comorbidities are frequent among type 1 diabetes patients on renal replacement therapy, yet the effect of comorbidities on survival is unknown. Our aim was to estimate this effect. METHODS: An incident cohort of all patients with type 1 diabetes entering chronic renal replacement therapy (n = 656) in Finland between 2000 and 2008 was followed until death or the end of follow-up on 31 December 2008. All data were obtained from the Finnish Registry for Kidney Diseases, which collects information on comorbidities at the start of renal replacement therapy. The main outcome measure was relative risk of death according to comorbidities. RESULTS: At start of renal replacement therapy, 22% of the patients with type 1 diabetes had coronary artery disease, 19% peripheral vascular disease, 11% cerebrovascular disease, 33% left ventricular hypertrophy and 7% heart failure. All these comorbidities were significant predictors of death in univariate analyses (RR 1.6-4.9). The 5 year survival probability of patients without comorbidities was 74%, while it was 56% and 37%, respectively, for those with one or more than one comorbidity. When the comorbidities were studied in a multivariate model, adjusting for age and sex, peripheral vascular disease (RR 1.9), left ventricular hypertrophy (RR 1.7) and heart failure (RR 2.5) remained independent risk factors for death. Calculations indicated that one-third of deaths in the study population could be attributed to comorbidities. CONCLUSIONS/INTERPRETATION: Among patients with type 1 diabetes entering renal replacement therapy, comorbidities are common and strong predictors of death. Therefore, it is essential to identify and adequately treat comorbidities.

Wegener's granulomatosis in Finland in 1981-2000: risk of dialysis-dependent renal disease.

OBJECTIVE: To determine the risk of renal insufficiency requiring transient or chronic dialysis and its contributing factors in patients diagnosed with Wegener's granulomatosis (WG) between 1981 and 2000 in Finland. PATIENTS AND METHODS: A retrospective cohort study using hospital discharge registers with a review of hospital case reports. Data were complemented with files from the Finnish Registry for Kidney Diseases. All files were reviewed by 8 October 2005. RESULTS: A total of 492 patients received a verified diagnosis of WG in 1981-2000. Of these, 84 (17%) needed dialysis by the end of follow-up. Of the 84 dialysed patients, 32 (38%) recovered initially, 32 (38%) needed chronic dialysis (dialysed > 3 months), and 19 (23%) received a kidney transplant. Forty-one (49%) of the dialysed patients were alive at the end of follow-up: 16 with a kidney transplant, 14 on dialysis, and 11 without dialysis. The cumulative rate of developing renal involvement leading to dialysis was 14.6% [95% confidence interval (CI) 11.6-18.2] at 5 years and 29.6% (95% CI 21.5-40.0) at 20 years after onset of WG symptoms. Elevated creatinine levels at presentation were associated with an increased rate of renal insufficiency requiring chronic dialysis. Age, gender, and involvement of any particular organ system had no significant effect. CONCLUSIONS: WG patients are at great risk of developing renal insufficiency, and this risk increases as the disease progresses. Elevated creatinine levels at presentation are associated with an increased risk of dialysis-dependent end-stage renal disease.

Risk factors for end-stage renal disease in a community-based population: 26-year follow-up of 25,821 men and women in eastern Finland.

BACKGROUND AND OBJECTIVE: There are very few European cohort studies assessing the risk factors of end-stage renal disease (ESRD) in a community-based population. This study investigated the predictors of ESRD in Finland. DESIGN: Prospective cohort study. SETTING: Eastern Finland. SUBJECTS: A random sample of 25,821 men and women aged 25-64 years from the national population register participating in three independent cross-sectional population surveys in 1972, 1977 and 1982. Only the subjects without diagnosis of ESRD or chronic kidney disease based on the national register data were included in the study. MAIN OUTCOME MEASURE: Initiation of renal replacement therapy (dialysis or kidney transplantation) identified from the Finnish Registry for Kidney Diseases through December 31, 2006. RESULTS: A total of 94 cases with ESRD were identified during a mean follow-up period of 26.5 years. In a multivariate proportional subdistribution hazard regression analysis, taking into account death as a competing risk event, diabetes (hazard ratio [HR] 4.76, 95% confidence interval [CI] 2.32-9.79), hypertension (HR 2.21, 95% CI 1.19-4.12), obesity defined as body mass index > or =30 kg m(-2) (HR 2.02, 95 %CI 1.10-3.71) and male gender (HR 1.68, 95% CI 1.19-4.12) were independent risk factors for ESRD. CONCLUSION: The findings of the present study confirm that modifiable risk factors play a major role in the development of ESRD in the North-European population. People with diabetes, hypertension or obesity should be considered as the target groups when planning preventive measures to control the future epidemic of ESRD.

Mortality due to bleeding, myocardial infarction and stroke in dialysis patients.

Essentials Mortality due to bleeding vs. arterial thrombosis in dialysis patients is unknown. We compared death causes of 201 918 dialysis patients with the general population. Dialysis was associated with increased mortality risks of bleeding and arterial thrombosis. Clinicians should be aware of the increased bleeding and thrombosis risks. SUMMARY: Background Dialysis has been associated with both bleeding and thrombotic events. However, there is limited information on bleeding as a cause of death versus arterial thrombosis as a cause of death. Objectives To investigate the occurrence of bleeding, myocardial infarction and stroke as causes of death in the dialysis population as compared with the general population. Methods We included 201 918 patients from 11 countries providing data to the ERA-EDTA Registry who started dialysis treatment between 1994 and 2011, and followed them for 3 years. Age-standardized and sex-standardized mortality rate ratios for bleeding, myocardial infarction and stroke as causes of death were calculated in dialysis patients as compared with the European general population. Associations between potential risk factors and these causes of death in dialysis patients were investigated by calculating hazard ratios (HRs) with 95% confidence intervals (CIs) by the use of Cox proportional-hazards regression. Results As compared with the general population, the age-standardized and sex-standardized mortality rate ratios in dialysis patients were 12.8 (95% CI 11.9-13.7) for bleeding as a cause of death (6.2 per 1000 person-years among dialysis patients versus 0.3 per 1000 person-years in the general population), 13.4 (95% CI 13.0-13.9) for myocardial infarction (22.5 versus 0.9 per 1000 person-years), and 12.4 (95% CI 11.9-12.9) for stroke (14.3 versus 0.7 per 1000 person-years). Conclusion Dialysis patients have highly increased risks of death caused by bleeding and arterial thrombosis as compared with the general population. Clinicians should be aware of the increased mortality risks caused by these conditions.

Past chlamydial infection is not associated with primary fallopian tube carcinoma.

We conducted a retrospective seroepidemiological study to evaluate the relationship between past chlamydial infection and primary fallopian tube carcinoma (PFTC). Postoperative serum samples were drawn from 79 consecutive patients treated for PFTC in 1985-2000. For each case two controls were selected. Serum samples were analysed for IgG antibodies to different C. trachomatis serotype pools and to C. pneumoniae. Seropositivity in general or serum antibody levels to different C. trachomatis serovars or C. pneumoniae did not differ between PFTC patients and controls. The lack of association between anti-chlamydial antibodies and PFTC suggests that past chlamydial infection does not play a role in the etiopathogenesis of PFTC. However, because chlamydial infection is common at young age and PFTC develops decades later, we cannot definitively exclude the possibility that C.trachomatis contributes to the development of PFTC.

Insulin-like growth factor I is not a useful marker of prostate cancer in men with elevated levels of prostate-specific antigen.

High serum levels of insulin-like growth factor I (IGF-I) and low levels of IGF-binding protein-3 (IGFBP-3) have been shown to correlate with increased prostate cancer risk. To evaluate this, IGF-I, IGFBP-3, and prostate-specific antigen (PSA) were measured in serum from 665 consecutive men (179 with prostate cancer), aged 55-67 yr, with elevated serum prostate-specific antigen (PSA; > or = 4 microg/L) in a screening trial. Men in the highest quartile of IGF-I levels had an odds ratio (OR) for prostate cancer of 0.50 [95% confidence interval (CI) 0.26-0.97] when adjusting for serum IGFBP-3. IGFBP-3 itself was not significantly associated with prostate cancer risk (OR, 1.24; 95% CI, 0.68-2.24). Prostate volume was larger in men without than in those with prostate cancer (P < 0.001), and after adjustment for prostate volume, the negative association between serum IGF-I and prostate cancer risk was no longer significant (OR, 0.57; 95% CI, 0.28-1.16). In screen-positive men with elevated serum PSA, serum IGF-I is not a useful diagnostic test for prostate cancer, but it may be associated with benign prostatic hyperplasia and enlargement.

Quantitative detection of low-copy-number mRNAs differing at single nucleotide positions.

Accurate analysis of mRNA expression levels of SNPs, highly homologous genes, and splicing variants requires techniques capable of quantifying low-copy-number mRNAs differing at single nucleotide positions. We have used an RT-PCR-based technique based on co-amplification of closely related target mRNA transcripts and assessed the effect of the stochastic distribution of low-copy-number templates on sampling variation when quantifying rare mRNA transcripts. The technique was optimized for maximal sensitivity to enable the analysis of samples containing a subpopulation of target cells and small microdissected samples. We demonstrate that the input level of template molecules is a critical determinant of the achievable assay precision. A minimum of approximately 50 molecules of template is required to discriminate between 2-fold differences in the expression levels of two transcripts. At levels above 1000 molecules of input template, the stochastic effects on sampling variation become negligible.

Antibiotic susceptibility of blood culture isolates after nearly two decades with netilmicin and ampicillin in neonatal septicaemia.

The aim of the study was to investigate the in vitro antibiotic susceptibility of blood culture isolates after almost 20 years with ampicillin and methicillin as empirical treatment for neonatal septicaemia. All blood culture isolates and their antibiograms obtained in a single tertiary neonatal intensive care unit from 1 January 1989 to 31 December 1994 were reviewed. Two hundred and six blood cultures from 181 infants containing 223 bacterial and 11 fungal isolates were identified during 4416 admissions. Fifteen (6.7%) of the bacterial isolates were resistant to ampicillin and netilmicin. Fourteen per cent of the staphylococcal spp. were susceptible to penicillin while more than 90% were susceptible to netilmicin. The coagulase-negative staphylococci (CONS) were resistant to netilmicin, methicillin and gentamicin in 12%, 49% and 65%, respectively. Eighty-nine per cent of the methicillin-resistant CONS were susceptible to netilmicin as opposed to 17% to gentamicin (p<0.001). Except for one strain of Acinetobacter sp., all Gram-negative bacteria were susceptible to netilmicin. Our data show that the ampicillin-netilmicin combination still provides a high in vitro coverage (93%) against bacteria identified in blood cultures from newborns in our unit. Netilmicin has a significantly better in vitro effectiveness against CONS than gentamicin.

C-reactive protein (CRP) response patterns in neonatal septicaemia.

C-reactive protein (CRP) is an unreliable diagnostic tool in the early diagnosis of neonatal septicaemia. However, serial measurements have been shown to be useful in monitoring the effectiveness of treatment. The aim of the present study was to investigate whether a specific CRP response pattern to different groups of pathogens could be identified during treatment of neonatal septicaemia. Serial CRP measurements from day 1 to 4 in monomicrobial blood culture-proven episodes of septicaemia were reviewed. In 4416 admissions, 180 out of 206 positive blood cultures were monomicrobial; 121 monomicrobial septic episodes were eligible for final analysis of the CRP response during treatment. A low median (M) value (day 1 to 4) was identified in coagulase-negative staphylococci (CONS) (M=23 mg/l), contrasting with high median values in Staphylococcus aureus (M=58 mg/l), group B streptococci (M=51 mg/l), Escherichia coli (M=51 mg/l) and Candida species (M=76 mg/l) (p<0.001). Median CRP values in the two groups were different for each of the treatment days 1 to 4 (p<0.001). An increase (p<0.001) in CRP during the 24 h before initiation of treatment was a sign of late-onset CONS septicaemia. In episodes where antimicrobial treatment failed, CRP levels were moderately elevated the day prior to treatment start and increased continuously thereafter, whereas successful treatment was generally accompanied by a decline in CRP in less than 4 days. The CRP response to CONS is significantly less pronounced than to other commonly encountered pathogens in neonatal septicaemia. A rise in CRP beyond the third day of empirical treatment should give rise to a suspicion of fungal infection or ineffective antibacterial treatment.

Induction of endometrial plasminogen activator-inhibitor 1: a possible mechanism contributing to the effect of intrauterine levonorgestrel in the treatment of menorrhagia.

OBJECTIVE: To elucidate molecular mechanisms accounting for excessive menstrual blood loss, and to present the therapeutic effect of an intrauterine levonorgestrel system (LNG-IUS) in menorrhagia. DESIGN: A multicenter study comparing hysterectomy with the LNG-IUS in treating menorrhagia. SETTING: A university hospital. PATIENT(S): Women with (n = 27) and without (n = 14) menorrhagia, and women with uterine fibroids but undetermined menstrual blood loss (n = 35). INTERVENTION(S): An LNG-IUS was inserted into the uterine cavity in 11 women with menorrhagia and six women experiencing normal menstrual blood loss. MAIN OUTCOME MEASURE(S): Expression of the messenger ribonucleic acid (mRNA) of tissue-type plasminogen activator (t-PA) and that of a specific PA inhibitor type 1 (PAI-1) in endometrial tissue samples, as evaluated with the use of Northern blot analysis. RESULT(S): t-PA mRNA was expressed in the endometrium throughout the menstrual cycle, with no statistically significant difference between a proliferative (n = 30) and a secretory endometrium (n = 40), or between women experiencing normal menstrual blood loss (n = 14) and those with menorrhagia (n = 27). The levels of t-PA mRNA in menstrual phase samples (n = 6) were significantly higher than those in proliferative or secretory endometrium. PAI-1 mRNA was detected in the endometrium during menstruation only. Both t-PA mRNA and PAI-1 mRNA were expressed in all endometrial samples (n = 17) obtained 6 months after an LNG-IUS was inserted, regardless of the menstrual cycle phase. The relative levels of both types of mRNA were significantly higher in LNG endometrium than in proliferative or secretory endometrium, but levels did not differ from those in menstrual-phase endometrium. CONCLUSION(S): The mean (+/-SD) levels of t-PA mRNA and PAI-1 mRNA in the endometrium of women with and without menorrhagia did not differ, suggesting that the PA system is not the major determinant of menstrual blood loss. However, continuous induction of PAI-1 may contribute to the therapeutic effect of LNG-IUS in treating menorrhagia.

Prevalence of 'do not attempt resuscitation' orders and living wills among patients suffering cardiac arrest in four secondary hospitals.

OBJECTIVE: To assess the prevalence and implementation of 'do not resuscitate' orders, nowadays called 'do not attempt resuscitation' (DNAR) orders and living wills among patients suffering in-hospital cardiac arrest (CA) in whom cardiopulmonary resuscitation was not initiated. MATERIALS AND METHODS: A prospective survey of CA patients conducted in four secondary hospitals during 2000-2001. The information collected included the presence of DNAR and a living will and the patients sociodemographic and disease factors and the reasons for not initiating resuscitation when no DNAR order was present. Data on the resuscitated patients were collected according to the Utstein recommendations (analyzed and published separately) and used for comparison. RESULTS: During the study period, 1486 patients suffered CA without resuscitation being initiated. Data collection was successful in 1143 patients (77%), who were included in the study. Most of the patients (84.5%) had a DNAR order. The prevalence of DNAR orders differed between the participating hospitals (P<0.001), and between the wards of the hospital, with most DNAR orders in the cardiac care unit (100%) and medical wards (87%). The patients designated as DNAR were likely to be older (P<0.01) and of poorer functional status (P<0.001). Reasons for abstaining from resuscitation without a DNAR order were unwitnessed arrest (27%) and terminal disease (66%). Living wills were uncommon (1.5%). Patients with a living will were likely to have a DNAR order (P<0.01). CONCLUSION: Most patients who suffered in-hospital CA without resuscitation had a DNAR order, and, for those who did not, terminal disease and medical futility were evident in most cases. Living wills were uncommon, but they appeared to have had some impact on treatment.

Evaluation of the in-hospital Utstein template in cardiopulmonary resuscitation in secondary hospitals.

INTRODUCTION: The in-hospital Utstein template for cardiopulmonary resuscitation (CPR) was assessed in four secondary hospitals (334-441 beds) which did not have systematic data collection. MATERIALS AND METHODS: The reports and outcome over a period of 12 months during the years 2000-2001 were evaluated. RESULTS: Of a total of 1690 patients that had a cardiac arrest (CA), 204 (12%) were resuscitated. Information on the collected Utstein parameters were available as follows: initial rhythm in 91%, time interval from collapse to defibrillation (in case of ventricular fibrillation or ventricular tachycardia as initial rhythm) in 90%, time interval to return of spontaneous circulation (ROSC) in 83% and duration of resuscitation in 83%. ROSC was achieved in 69 patients (34%, CI 27-41%) and 34 (17%, CI 11-23%) survived to hospital discharge. Twenty patients showed satisfactory neurological recovery (10%, CI 6-14%). Eighteen (9%, CI 5-13%) patients were alive at 12 months from the event. Factors associated with survival to hospital discharge were VF/VT (P=0.007) as the initial rhythm and shorter interval to defibrillation (P=0.046). CONCLUSION: The in-hospital Utstein template was logical but laborious and it provided tools for resuscitation management evaluation in the study hospitals. For continuous use, a slightly compressed model may be warranted. In the present material, the overall survival rate to hospital discharge was in line with previous reports but there were somewhat less neurologically satisfactory survivors. There is an evident need to improve the outcome of patients suffering CA on the wards. An important step is to reduce the time interval to defibrillation.

Factors associated with progression of IgA nephropathy are related to renal function--a model for estimating risk of progression in mild disease.

BACKGROUND: A number of factors are linked to the outcome of IgA nephropathy (IgAN). However, it has been difficult to compare results of studies since patient populations have varied greatly. There were 3 aims in the study reported here, namely to compare factors associated with renal outcome in IgAN patients with different levels of renal function on diagnosis; to determine factors which were independently associated with progression of renal disease in initially mild IgAN; and to create a model for the estimation of the risk of progression in individual IgAN patients with normal renal function on diagnosis. METHODS: Two hundred and fifty-nine IgAN patients who had been followed on average for 9.1 (SD 4.5) after diagnosis were divided into 2 groups on the basis of renal function on diagnosis. In group 1 (98 patients), Ccr (creatinine clearance, estimated by the Cockcroft-Gault formula) was < 85 ml/min, in group 2 (161 patients) > or = 85 ml/min. Univariate analyses were used to find significant differences between progressors and non-progressors in both groups. Logistic regression analysis was used to determine factors independently associated with progression in group 2. RESULTS: Several factors were found to be associated with outcome in both groups, such as hypertension, level of Ccr, serum cholesterol, proteinuria, and also histopathological changes. Factors associated with progression in patients with initially decreased renal function (group 1), were predictable, such as male sex, absence of episodes of macroscopic hematuria, serum urate level and degree of tubular atrophy. Surprisingly, in patients with initially normal renal function (group 2), numbers of urinary erythrocytes were associated with outcome. The factors independently associated with progression in this group were number of urinary erythrocytes, existence of hypertension and in histopathology arteriolosclerosis and the level of glomerular score. A model for estimating risk of progression on the basis of various combinations of factors found to be independently associated with outcome is presented. CONCLUSIONS: We concluded that association between variable and outcome in IgAN depends partly on renal function at the time of assessment of the factor. Since there are factors which are independently associated with the outcome of early and apparently mild disease, early diagnosis of IgAN is desirable: outcome in mild IgAN can be predicted reliably on the basis of factors found to be independently associated with outcome.

Characterization and determination of the complex between prostate-specific antigen and alpha 1-protease inhibitor in benign and malignant prostatic diseases.

Prostate-specific antigen (PSA) is a tissue-specific serine protease which forms complexes with protease inhibitors such as alpha 1-antichymotrypsin and alpha 2-macroglobulin. We have studied the interaction between PSA and alpha 1-protease inhibitor (API) in vitro and found that 15% of the added PSA binds to API while the majority of API is cleaved between Met358 and Ser359 when PSA is incubated with a 5-fold excess of API at 37 degrees C for 7 days. The complex between PSA and API (PSA-API) formed in vitro displays the same chromatographic behavior, molecular size and immunoreactivity as endogenous PSA-API occurring in serum, indicating that they are identical. PSA-API can be detected in serum by a time-resolved immunofluorometric assay (IFMA), in which a monoclonal antibody to PSA is used as a catcher and a polyclonal antibody to API labeled with a Eu-chelate is used as a tracer. Purified PSA-API formed in vitro is used as a calibrator. PSA-API in serum represents 1.0-7.9% (median 2.4%) of total PSA (tPSA) in prostate cancer (PCa, n = 82) and 1.3-12.2% (median 3.6%, p < 0.01) in patients with benign prostatic hyperplasia (BPH, n = 66). The IFMA for PSA-API in serum is hampered by a variable background, which is caused by non-specific adsorption of the huge excess of API in serum to the solid phase. The background can be determined by an assay using the same tracer as in the IFMA for PSA-API but PSA-unrelated antibody on the solid phase. The background signal is subtracted from the PSA-API signal. The clinical utility of PSA-API in serum has been evaluated in PSA-positive subjects from the Finnish PCa screening trial. After subtraction of the background, the proportion of PSA-API in relation to tPSA is lower in PCa than in controls, 0.9% vs. 1.6%, respectively (p < 0.001). Logistic regression analysis showed that the concentration of PSA-API was independent of the proportion of free PSA as a diagnostic variable among subjects with a tPSA of 4-10 micrograms/l (p = 0.009). The probability of PCa calculated by logistic regression using the concentration of PSA-API and the proportion of free PSA in serum significantly improved cancer specificity at high sensitivity levels (85-95%) as compared to the proportion of free PSA alone.

Mortality due to pulmonary embolism, myocardial infarction, and stroke among incident dialysis patients.

BACKGROUND: It is has been suggested that dialysis patients have lower mortality rates for pulmonary embolism than the general population, because of platelet dysfunction and bleeding tendency. However, there is limited information whether dialysis is indeed associated with a decreased mortality risk from pulmonary embolism. OBJECTIVE: The aim of our study was to evaluate whether mortality rate ratios for pulmonary embolism were lower than for myocardial infarction and stroke in dialysis patients compared with the general population. METHODS: Cardiovascular causes of death for 130,439 incident dialysis patients registered in the ERA-EDTA Registry were compared with the cardiovascular causes of death for the European general population. RESULTS: The age- and sex-standardized mortality rate (SMR) from pulmonary embolism was 12.2 (95% CI 10.2-14.6) times higher in dialysis patients than in the general population. The SMRs in dialysis patients compared with the general population were 11.0 (95% CI 10.6-11.4) for myocardial infarction, 8.4 (95% CI 8.0-8.8) for stroke, and 8.3 (95% CI 8.0-8.5) for other cardiovascular diseases. In dialysis patients, primary kidney disease due to diabetes was associated with an increased mortality risk due to pulmonary embolism (HR 1.9; 95% CI 1.0-3.8), myocardial infarction (HR 4.1; 95% CI 3.4-4.9), stroke (HR 3.5; 95% CI 2.8-4.4), and other cardiovascular causes of death (HR 3.4; 95% CI 2.9-3.9) compared with patients with polycystic kidney disease. CONCLUSIONS: Dialysis patients were found to have an unexpected highly increased mortality rate for pulmonary embolism and increased mortality rates for myocardial infarction and stroke.

Senescence rates in patients with end-stage renal disease: a critical appraisal of the Gompertz model.

The most frequently used model to describe the exponential increase in mortality rate over age is the Gompertz equation. Logarithmically transformed, the equation conforms to a straight line, of which the slope has been interpreted as the rate of senescence. Earlier, we proposed the derivative function of the Gompertz equation as a superior descriptor of senescence rate. Here, we tested both measures of the rate of senescence in a population of patients with end-stage renal disease. It is clinical dogma that patients on dialysis experience accelerated senescence, whereas those with a functional kidney transplant have mortality rates comparable to the general population. Therefore, we calculated the age-specific mortality rates for European patients on dialysis (n=274 221; follow-up=594 767 person-years), for European patients with a functioning kidney transplant (n=61 286; follow-up=345 024 person-years), and for the general European population. We found higher mortality rates, but a smaller slope of logarithmic mortality curve for patients on dialysis compared with both patients with a functioning kidney transplant and the general population (P<0.001). A classical interpretation of the Gompertz model would imply that the rate of senescence in patients on dialysis is lower than in patients with a functioning transplant and lower than in the general population. In contrast, the derivative function of the Gompertz equation yielded the highest senescence rates for patients on dialysis, whereas the rate was similar in patients with a functioning transplant and the general population. We conclude that the rate of senescence is better described by the derivative function of the Gompertz equation.

Effect of general population mortality on the north-south mortality gradient in patients on replacement therapy in Europe.

In Europe there is considerable variation in mortality on renal replacement therapy (RRT). The causes of this variation are still poorly understood. We hypothesized that differences in mortality in the general population contribute to differences in mortality on RRT. To evaluate this relationship, we studied general population statistics obtained from Eurostat and the individual data of 67,692 patients on RRT from 15 national and regional renal registries. These 15 registries were divided into two geographical regions: North and South Europe. Cox regression was used to assess the relative risk of death (RR) for each region with adjustment for age, gender, diabetes, and additionally general population mortality. In patients on RRT the age, gender and diabetes adjusted RR of death was 0.65 (95% CI (0.64-0.66)) for South compared to North, while in the general population the age and gender standardized RR of death was 0.91. After adjustment for general population mortality in addition to age, gender, and diabetes, the RR of death for patients on RRT in the South changed from 0.65 to 0.74 (95% CI (0.72-0.75)), which indicates that general population mortality accounted for 26% of the region-related mortality difference on RRT. In conclusion, within Europe there exist considerable international differences in the mortality of patients on RRT. Twenty-six percent of the European north-south mortality difference in RRT could be attributed to differences in general population mortality. Our data support the hypothesis that general population mortality is an important factor to take into account when making RRT mortality comparisons.