Summary
Background and objectives
CKD and ESRD are growing burdens. It is unclear whether these conditions affect pulmonary embolism (PE) risk, given that they affect both procoagulant and anticoagulant factors. This study examined the frequency and associated outcomes of PE in CKD and ESRD.
Design, setting, participants, & measurements
The Healthcare Cost and Utilization Project’s Nationwide Inpatient Sample was used to estimate the frequency and outcomes of PE in adults with CKD and ESRD. Hospitalizations for the principal diagnosis of PE and presence of CKD or ESRD were identified using International Classification of Diseases, Ninth Revision codes. Data from the annual US Census and US Renal Data System reports were used to calculate the number of adults with CKD, ESRD, and normal kidney function (NKF) as well as the annual incidence of PE in each group. Logistic regression modeling was used to compare in-hospital mortality among persons admitted for PE who had ESRD or CKD to those without these conditions.
Results
The annual frequency of PE was 527 per 100,000, 204 per 100,000, and 66 per 100,000 persons with ESRD, CKD, and NKF, respectively. In-hospital mortality was higher for persons with ESRD and CKD (P<0.001) compared with persons with NKF. Median length of stay was longer by 1 day in CKD and 2 days in ESRD than among those with NKF.
Conclusions
Persons with CKD and ESRD are more likely to have PE than persons with NKF. Once they have PE, they are more likely to die in the hospital.
Introduction
CKD and ESRD are common and costly conditions in the United States (1,2). Estimates using the US Renal Data System (USRDS) suggest that in 2010 approximately 20 million Americans aged ≥20 years had CKD; 0.39 million had ESRD, excluding persons with renal transplants (3–6). Persons with these conditions are frequently hospitalized and incur greater costs and complications associated with their hospitalizations (7).
The overall incidence of pulmonary thromboembolism (PE) in adults is estimated to be 1 event per 1000 person-years (8,9). Although the case fatality rate for PE has fallen below 10%, it remains an important cause of death in hospitalized patients, leading to approximately 300,000 deaths annually (8–10).
Although CKD is thought to be a procoagulant state, associated platelet dysfunction is thought to enhance bleeding risk, particularly as CKD progresses to ESRD. Most studies have shown that persons with CKD have an increased risk of venous thromboembolism (VTE) (11–14), but estimates of the magnitude of the effect vary widely. Moreover, prior investigations of the risk of VTE and/or PE among persons with ESRD show conflicting results (15–18). Thus, more research is needed to understand the risk of VTE, and especially PE, among persons with renal disease. This is also important because the presence of renal disease increases the risks associated with administration of intravenous contrast (a potentially nephrotoxic agent) and limits use of low molecular weight heparins. This may lead to delayed diagnosis, prolonged hospital stays, and increased costs when such persons have VTE.
We therefore carried out this study to examine the frequency of PE among persons with either CKD or ESRD compared with persons with normal kidney function. In addition, we describe the effect of CKD and ESRD on length of stay and in-hospital mortality among persons who are hospitalized with PE. We used data from a large nationally representative hospital database to enhance the external validity of our results.
Materials and Methods
Data Source
We used the Healthcare Cost and Utilization Project's Nationwide Inpatient Sample (NIS), the largest publicly available all-payer database regarding inpatient care in the United States. This is an administrative dataset created by the Agency for Healthcare Research and Quality from data contributed by participating states. Each year, the NIS includes data on 5–8 million hospital stays from approximately 1000 hospitals selected to approximate a 20% stratified sample of US hospitals. All hospital types are sampled, excluding Federal (e.g., Veterans Affairs or Department of Defense), institutional (e.g., prison hospitals), and short-term rehabilitation hospitals. Each hospitalization is treated as an individual entry in the database and is coded with 1 principal diagnosis, up to 14 secondary diagnoses, and 15 procedural diagnoses associated with that stay. To facilitate the production of national estimates, both hospital and discharge weights are provided. Details about the structure of NIS database are available online (19). We used data from 2007 for this study.
Study Population
We used the NIS to estimate the total number of admissions for PE among adult patients (aged ≥20 years), and then classified such admissions into three distinct cohorts of persons with CKD, ESRD, or normal kidney function. We included all admissions with a principal diagnosis code for PE (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 415.11 and 415.196), indicating that PE was the primary reason for admission (20). We then used ICD-9-CM codes to classify each patient as having CKD, ESRD, or normal kidney function. We identified CKD patients using ICD-9-CM codes 585.1–585.5 and 585.9. Similarly, we considered patients to have ESRD if they had either the diagnosis code of 585.6 or a procedure code for hemodialysis (39.95) or peritoneal dialysis (54.98). We excluded persons undergoing hemodialysis who concomitantly had an ICD-9-CM code indicating ARF. We also excluded patients with renal transplantation from our analysis. Figure 1 details our scheme of patient selection and inclusion.
Figure 1.
Selection of ESRD and CKD hospitalizations for persons aged ≥20 years. ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification.
We identified known clinical risk factors for PE using ICD-9-CM codes and described the proportion of persons with these risk factors in each cohort studied. The risk factors identified were congestive heart failure (428), diabetes mellitus (250), coronary artery disease (410, 411), obesity (278.00–278.02), smoking (305.1, V15.82), cancer (140–175, 179–208), hypercoagulable states (270.4, 289.81), cirrhosis (571, 572), stroke (433, 434, 436, 437.1), hypertension (401–405), pregnancy (630–677), and chronic obstructive pulmonary disease (490–492).
Definition of Independent Variables
We used Deyo's modification of Charlson's comorbidity index to identify burden of comorbid disease (21). This index uses 17 comorbid conditions with differential weighting; scores range from 0 to 33, with higher scores representing greater comorbid burden. We excluded CKD or ESRD in calculating the index because they were already included as patient characteristics. We considered patients to have severe PE if they had identifying ICD-9-CM codes for the use of mechanical ventilation (96.7 or 96.72), vasopressors (00.17), or thrombolytic agents (99.10) on the day of admission or the second hospital day. We identified frequency of inferior vena cava (IVC) filter placement using ICD-9-CM code 38.7. We considered patients with CKD or normal kidney function to undergo new dialysis if they had an ICD-9-CM procedure code of hemodialysis (39.95).
Outcomes
Our primary outcome of interest were the frequency of pulmonary embolism, in-hospital mortality, lengths of hospital stay, and associated charges among persons admitted with PE. We compared each of these outcomes variables in the cohort of CKD and ESRD separately with the cohort of persons with normal kidney function. We also examined complications associated with PE (cardiogenic shock, respiratory failure requiring mechanical ventilation, and cardiac arrest) using appropriate ICD-9-CM codes.
Statistical Analyses
We estimated the number of US adults with CKD by multiplying the number of adults in age categories (20–44, 45–54, 55–64, 65–74, and ≥75 years) by the age-specific prevalence of CKD in each of these age groups. We estimated the prevalence of CKD among persons aged <65 years by averaging the estimates generated by analyses of the Ingenix and MarketScan databases published in the USRDS 2007 annual report (22). We used the estimates generated by analyses of Medicare databases published in the USRDS 2007 annual report to estimate the prevalence of CKD among persons aged ≥65 years. We used the number provided by the USRDS as our estimate of the number of Americans with ESRD in each of these age groups (3). We calculated the number of people with normal kidney function by subtracting our estimates of the number with ESRD or CKD in each age group from the estimated US population in these age groups from the US Census (23).
We used the strata weights provided with the NIS and SURVEY commands in the Stata IC 11.0 software program (Stata Corp, College Station, TX) to generate national estimates of the number of admissions for PE among adults with CKD, ESRD, or normal kidney function. We then calculated the annual incidence of PE for each age group for each category (normal kidney function, CKD, or ESRD) by dividing the number of admissions for PE by our estimate of the total US population in that category. We calculated an overall incidence of PE in each renal disease category in the same way. We then standardized the incidence of PE in the CKD and ESRD categories to the age distribution of the normal US population.
Among persons admitted with PE, we compared the distribution of outcomes and potential confounders among persons with CKD, ESRD, or normal kidney function using chi-squared tests to compare categorical variables and either t tests or Wilcoxon signed rank sum tests to compare continuous variables as appropriate for their distribution. Variables in both the CKD and ESRD cohorts were compared separately with the cohort of persons with normal kidney function. All tests were two sided and differences were deemed significant at P<0.05.
We constructed a multivariable logistic regression model to assess the independent effect of CKD and ESRD on in-hospital mortality in persons suffering PE. First we identified each characteristic associated with in-hospital mortality in univariate analysis at P<0.10. These were candidates for inclusion in the final multivariable model. We checked these variables for multicollinearity using tolerance and a variation inflation factor; these were very close to unity for all variables used in the final model. To account for interactions between selected variables, we examined all two-way interaction terms and included those found significant in the model. Information regarding race was missing in approximately 20% of patients and we grouped these together as unknown. We adjusted the final model for age (20–44, 45–54, 55–64, 65–74, and ≥75 years), sex, race (white, black, Hispanic, Asian, Native American, other, and unknown), modified Charlson–Deyo’s comorbidity index, and hospital characteristics such as hospital teaching status (e.g., teaching and nonteaching) and bed size (e.g., small, medium, and large). We excluded persons with missing data regarding age, sex, or discharge status. We used Stata IC 11.0 software for all analysis. This study was approved by Institutional Review Board of the Medical College of Wisconsin.
Results
We estimated that there were 32,616,411 adult discharges from hospitals covered by the NIS 2007 database, 154,585 of whom had PE. Of these, 143,060 had normal kidney function, 9920 had CKD, and 1605 had ESRD. On the basis of US Census and USRDS data, we calculated that among the 218,837,849 estimated adults (aged ≥20 years) living in the United States in 2007, 214,567,689 had normal kidney function, 3,757,005 had CKD, and 362,512 had ESRD. Adjusting for differences in age distribution between these groups and normalizing to the entire 2007 US population with normal kidney function, we calculated that the overall number of admissions for the principal diagnosis of PE in 2007 was 66 per 100,000 among persons with normal kidney function, 204 per 100,000 among persons with CKD, and 527 per 100,000 among persons with ESRD. The frequency of PE by age group and renal disease category is shown in Table 1.
Table 1.
Age standardized frequency of pulmonary embolism per 100,000 persons in 2007 by kidney disease category
| Age Group (yr) | Pulmonary Embolism Admissions per 100,000 Persons by Kidney Disease Category | ||
|---|---|---|---|
| Normal Kidney Function | CKD | ESRD | |
| 20–44 | 24 | 137 | 682 |
| 45–54 | 51 | 236 | 367 |
| 55–64 | 80 | 312 | 312 |
| 65–74 | 154 | 220 | 456 |
| ≥75 | 249 | 304 | 462 |
| All | 66 | 204 | 527 |
Demographic and Clinical Characteristics
Persons with PE in the ESRD cohort were significantly younger compared with persons with normal kidney function. Conversely, persons with CKD were significantly older. Even after removing CKD as a comorbid condition, both persons with CKD and those with ESRD had a greater Charlson–Deyo index and were more likely to have individual comorbid conditions such diabetes mellitus, hypertension, and congestive heart failure than persons with normal kidney function. The proportion of PE admissions that we categorized as severe was highest in the ESRD cohort followed by the CKD cohort; both were significantly higher than among persons with normal kidney function (Table 2).
Table 2.
Demographic and clinical characteristics of pulmonary embolism admissions according to kidney function
| Normal Kidney Function | CKD | ESRD | |
|---|---|---|---|
| Pulmonary embolism (n) | 142,338 | 9920 | 1600 |
| Age (yr), mean ± SD | 62.3±17.3 | 73.1±13.3a | 59.6±17.2a |
| Age (yr), n (%) | |||
| 20–44 | 24,766 (17.4) | 397 (4.0)a | 361 (22.6)a |
| 45–54 | 22,191 (15.6) | 605 (6.1)a | 232 (14.5) |
| 55–64 | 25,819 (18.1) | 1328 (13.4)a | 272 (17.0) |
| 65–74 | 28,113 (19.8) | 2370 (23.9)a | 374 (23.4) |
| ≥75 | 41,447 (29.1) | 5220 (52.6)a | 360 (22.5)a |
| Sex, n (%) | |||
| Male | 63,791(44.8) | 5188 (52.3) | 713(44.5) |
| Race, n (%) | |||
| White | 76,731 (53.9) | 5202 (52.4) | 465 (29.1)a |
| Black | 15,793 (11.1) | 1470 (14.8) | 532 (33.2)a |
| Hispanic | 5226 (3.6) | 330 (3.3) | 67 (4.1) |
| Asian | 860 (0.6) | 77 (0.8) | 29 (1.8)a |
| Native American | 502 (0.4) | 31 (0.3) | 19 (1.2)a |
| Other | 2083 (1.5) | 91 (0.9) | 30 (1.9) |
| Unknown | 41,143 (28.9) | 2718 (27.4) | 457 (28.5) |
| Charlson–Deyo's comorbidity score, n (%) | |||
| 0 | 63,387 (44.5) | 2806 (28.2)a | 457 (28.5)a |
| 1–2 | 57,256 (40.2) | 5190 (52.3)a | 811 (50.7)a |
| 3–4 | 8426 (5.9) | 1294 (13.0)a | 207 (12.9)a |
| 5–6 | 2912 (2.0) | 148 (1.5) | 58 (3.6) |
| ≥7 | 10,355 (7.3) | 480 (4.8) | 67 (4.2) |
| Severe pulmonary embolism, n (%) | 3951 (2.8) | 429(4.3)a | 89 (5.5)a |
| Comorbidities (%) | |||
| Diabetes mellitus | 18.1 | 34.2a | 38a |
| Hypertension | 49.6 | 74.5a | 92.5a |
| Congestive heart failure | 10.0 | 43.1a | 33.3a |
| Morbid obesity | 5.6 | 5.5 | 4.0 |
| Smoking | 20.5 | 11.5a | 8.8a |
| Chronic obstructive pulmonary disease | 15.8 | 24.2a | 17.0 |
| Cancer | 23.5 | 21.3 | 13.2a |
| Hypercoagulable state | 3.9 | 2.0a | 1.8a |
| Cirrhosis | 1.5 | 1.1 | 1.3 |
Significant difference at P<0.05 between the variable compared with normal kidney function using the chi-squared test.
Use of Thrombolysis and IVC Filters
The proportion of patients undergoing thrombolysis was similar in all categories. However, IVC filters were more commonly placed in persons with CKD (15.3%) compared with those with ESRD (13.6%) or normal kidney function (13.1%) (Table 3).
Table 3.
Outcomes of pulmonary embolism hospitalizations among persons with CKD, ESRD, or normal kidney function
| Normal Kidney Function | CKD | ESRD | |
|---|---|---|---|
| Outcomes | |||
| Mortality, n (%) | 4523 (3.2) | 666 (6.7)a | 106 (6.7)a |
| LOS (survivor days), median (IQR) | 5 (3–7) | 6 (4–8)b | 7 (4–10)b |
| Disposition of survivors, n (%) | |||
| Home | 89,298 (64.7) | 4606 (49.8) | 941 (63) |
| Nursing home | 21,184 (15.4) | 2260 (24.4) | 253 (16.9) |
| Home care | 23,245 (16.9) | 2146 (23.2) | 205 (13.7) |
| Other hospitals | 3106 (2.3) | 203 (2.2) | 58 (3.9) |
| Otherc | 979 (0.7) | 37 (0.4) | 36 (2.4) |
| Procedures (%) | |||
| Thrombolysis | 1.6 | 1.7 | 2.6 |
| IVC filter | 13.1 | 15.3a | 13.6 |
| Complications (%) | |||
| Mechanical ventilation | 2.2 | 4.6a | 5.6a |
| Mechanical ventilation >96 h | 0.7 | 1.7a | 1.9a |
| Cardiac arrest | 0.8 | 1.5a | 2.6a |
| New dialysis | 0.1 | 0.9a | — |
LOS, length of stay; IQR, interquartile range; IVC, inferior vena cava.
P<0.05 compared with normal kidney function using the chi-squared test.
P<0.05 compared with normal kidney function using the Wilcoxon signed rank test.
Includes against medical advice and unknown.
Outcomes
The unadjusted in-hospital mortality in patients with PE was 3.2% in patients with normal kidney function. Mortality was significantly higher for persons with CKD (6.7%) and ESRD (6.7%). On adjusted analysis using multivariable regression, the odds ratio for mortality in CKD patients was 1.57 (95% confidence interval, 1.27–1.93) compared with those with normal kidney function (Table 4). Similarly the odds of mortality in persons with ESRD suffering PE were 1.92 times higher (95% confidence interval, 1.17–3.15) compared with normal kidney function. Table 4 demonstrates the multivariable analysis of predictors of mortality in PE. The proportion of CKD patients with PE requiring new dialysis was 0.9%, which was nine times higher than among persons with normal kidney function (Table 3).
Table 4.
Logistic regression analysis of predictors associated with increased mortality in pulmonary embolism
| Odds Ratio (95% Confidence Interval) | |
|---|---|
| Renal function | |
| Normal kidney function | Reference |
| CKDa | 1.57 (1.27–1.93) |
| ESRDa | 1.92 (1.17–3.15) |
| Age (yr) | |
| 20–44 | Reference |
| 45–54a | 1.59 (1.13–2.23) |
| 55–64a | 2.13 (1.55–2.91) |
| 65–74a | 2.71 (2.00–3.69) |
| ≥75a | 4.60 (3.44–6.16) |
| Sex | |
| Male | Reference |
| Female | 1.03 |
| Race | |
| White | Reference |
| Blacka | 1.30 (1.06–1.59) |
| Hispanic | 1.19 (0.85–1.65) |
| Asian | 0.62 (0.24–1.58) |
| Native American | 1.19 (0.42–3.31) |
| Other | 1.45 (0.86–2.46) |
| Unknown | 1.06 (0.91–1.24) |
| Modified Charlson–Deyo's comorbidity index | |
| 0 | Reference |
| 1–2a | 1.69 (1.43–1.99) |
| 3–4a | 2.05 (1.59–2.65) |
| 5–6a | 4.66 (3.30–6.59) |
| ≥7a | 5.81 (4.72–7.16) |
| Severe pulmonary embolisma | 22.6 (19.1–26.9) |
| Hospital teaching status | |
| Nonteaching hospital | Reference |
| Teaching hospital | 1.06 (0.93–1.21) |
| Hospital bed size | |
| Small | Reference |
| Mediuma | 0.76 (0.61–0.96) |
| Large | 0.95 (0.77–1.15) |
P<0.05.
Length of Hospital Stay and Disposition
Median length of hospital stay in survivors was 2 days longer for ESRD patients and 1 day longer for CKD patients compared with persons with normal kidney function (P<0.001). The numbers of patients with PE discharged to a health care facility were significantly higher in those with CKD or ESRD. Discharge to home was lowest for persons with CKD suffering PE.
Discussion
We report that admission for PE is more frequent among persons with ESRD or CKD than among those with normal kidney function. PE in persons with ESRD and CKD was also associated with a significantly higher case fatality rate. This effect persisted and was significant after adjustment for potential confounding variables. These results emphasize the need to consider prophylaxis, diagnosis, and treatment of PE in persons with renal disease, despite the increased risk of complications.
Hemostasis is an intricately balanced process. Abnormalities in platelet function and the coagulation cascade may lead to either increased risk of clotting or bleeding, depending on the defect. Mechanisms contributing to a procoagulant state in persons with CKD include increased tissue factor, vWf, XIIa, VIIa, and fibrinogen levels along with reduced tissue plasminogen activator (24–30). With progression to ESRD, patients develop platelet dysfunction that may lead to enhanced bleeding risk (31–35). Conversely, increased levels of procoagulant factors such as homocysteine, thrombin-antithrombin complex, D-dimer, fibrinogen, vWf, and protein C deficiency in ESRD patients may increase the prothrombotic state (36–38). Patients with ESRD or CKD may be more likely to have risk factors for PE such as congestive heart failure and immobility, in addition to the aforementioned molecular mechanisms. Use of erythropoietin in ESRD patients has also been associated with an elevated risk of thrombosis (39). Thrombophilic factors are also common in stage 5 CKD and correctable after kidney transplantation (40).
Our finding that the incidence of PE in persons with ESRD is 7.6 times higher than in persons with normal kidney function is similar to that of Tveit et al.; however, their estimate of the incidence ratio was slightly lower at 6.09 (15). This may reflect differences in methodology. Tveit et al. estimated the incidence of PE in ESRD patients using the USRDS database and that in the general US population using the National Hospital Discharge Survey database; conversely, we used a single database (NIS) to estimate the frequency of PE in both cohorts. They also restricted their ESRD cohort to persons in their first year after initiation of dialysis, whereas our sample included all persons receiving dialysis, some of whom would have had ESRD for many years. Finally, although Tveit et al. included only the first episode of PE to identify persons with PE in the ESRD cohort, they used all hospitalizations for PE to estimate incidence in the general population. We only included persons admitted with a principal diagnosis of PE and used the same technique to identify cases among each of our study groups (those with normal kidney function, CKD, and ESRD).
Autopsy studies have suggested that persons with ESRD have a reduced risk of PE (17,41,42). In one large series, Mossey et al. reported microscopic or macroscopic PE to be present in 32.3% of patients with normal kidney function, whereas the frequency of PE in patients with stage 5 CKD or ESRD was 9.5%, with all of them being microscopic emboli (41). More recently, Daneschvar et al. found no difference in VTE prevalence between autopsied persons with versus without CKD (43). These postmortem studies are limited by selection bias. Significantly higher mortality in persons with ESRD secondary to infectious and other cardiovascular complications may spuriously lower the frequency of PE on autopsy when compared with the general population. Furthermore, the subset of persons undergoing autopsy examinations is relatively small and the clinical significance of microscopic thrombotic events discovered postmortem is unclear. We have attempted to circumvent such issues by using the USRDS database to determine the true at-risk ESRD and CKD population and including persons with the principal diagnosis of PE only. We judged this approach better than including any diagnosis of PE to prevent spurious overinflation of PE in the numerator.
Although they excluded persons with ESRD, our results are qualitatively similar to those obtained by Folsom et al. (13), as well as those by Wattanakit et al. in their analysis of data from the Longitudinal Investigation of Thromboembolism Etiology (LITE) (11). However, the magnitude of the effect of CKD on risk of VTE in both of these studies was less than we observed. This likely reflects differences in methods. We estimated the US population with CKD during the same year we identified the number of admissions for PE, whereas they defined CKD at baseline then identified VTE several years later. Thus, the CKD status of many in the LITE cohort likely changed by the time VTE incidence was measured. This might attenuate any effect of CKD on VTE risk. Furthermore, we examined PE only, whereas they measured any VTE.
Our finding of higher mortality in patients with CKD and ESRD suffering PE is consistent with previous reports (44). We also report higher frequency of severe PE for both the CKD and ESRD populations compared with persons with normal kidney function. We postulate that increased comorbidity burden, especially cardiovascular complications associated with both these diagnoses, lead to diminished cardiopulmonary reserve, increasing both severity of PE and associated mortality.
Although our study uses a well characterized database, we must acknowledge some important limitations. Despite its national scope, we may not reliably identify the diagnoses of CKD, which is one of the inherent limitations of administrative database studies. We cannot exclude variations in coding practices among hospitals. Because the NIS is a discharge-level database, differentiation between multiple hospitalizations by the same patient is not possible; this may slightly exaggerate the precision of our estimates. The limited clinical detail of our ICD-9-CM codes cannot detect various stages of CKD, thereby forcing us to forgo reporting the incidence of PE in different stages of CKD. Furthermore, we calculated the number of persons in each renal disease category using US Census and USRDS annual reports. Whereas the total ESRD patients in United States are accurately counted, the same may not be true for CKD. Our estimate of PE incidence in persons with CKD would be exaggerated if the USRDS data underestimate the number of persons with CKD. Furthermore, we only included patients with a principal diagnosis of PE. Thus, our analysis does not include persons who developed PE after being admitted to the hospital for another diagnosis. Although we used multivariable regression to control for potential confounding factors, residual bias resulting from unmeasured severity of illness cannot be excluded.
Despite these limitations, our study for the first time provides an estimate of the frequency and mortality of PE admissions in both the CKD and ESRD population. We show that patients with ESRD and CKD not only have higher incidence of PE, but also develop more severe disease and have higher mortality. These findings have implications for the diagnosis, treatment and, importantly, prevention of PE in these cohorts. Definitive evaluations in the modern era require administration of intravenous contrast agents for computed tomography angiography. In persons with CKD, such an approach requires prophylaxis with intravenous fluids before contrast exposure and carries the risk of significant nephrotoxicity. This creates a layer of complexity and may delay diagnosis in a disease state in which prompt recognition and treatment is paramount. Conversely, there is the unnecessary risk of contrast-induced kidney injury in patients with negative studies, which may itself lead to increased lengths of hospital stay, mortality, and costs. In addition, therapy with low molecular weight heparin once a diagnosis is established is not feasible in persons with CKD with an estimated GFR <30 or in persons with ESRD. Thus, therapy would require hospitalization for administration of intravenous heparin and initiation of oral anticoagulant therapy until anticoagulation targets are achieved, thereby leading to increased costs. Furthermore, given the increased risk of death from PE, future investigations should evaluate current deep vein thrombosis prophylaxis strategies for these cohorts of patients. In summary, although persons with CKD and ESRD have both higher risk of PE and mortality from PE, risks associated with evaluation and associated costs of hospitalizations may be higher as well. Investigations focusing on cost-benefit analyses and determining the risk-benefit ratios of evaluating and treating persons with CKD and ESRD with PE are warranted.
Disclosures
None.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
References
- 1.US Renal Data System: Incidence and prevalence. In: Atlas of End-Stage Renal Disease Available at: http://www.usrds.org/2009/pdf/V2_02_INC_PREV_09.pdf Accessed November 10, 2011
- 2.US Renal Data System: Chronic kidney disease in the adult NHANES population. In: Atlas of Chronic Kidney Disease Available at: http://archive.usrds.org/2009/pdf/V1_01_09.PDF Accessed November 10, 2011
- 3.US Renal Data System: RenDER. Available at: http://www.usrds.org/render/xrender.phtml Accessed March 3, 2012
- 4.US Centers for Disease Control and Prevention: National Chronic Kidney Disease Fact Sheet, 2010. Available at: http://www.cdc.gov/diabetes/pubs/pdf/kidney_Factsheet.pdf Accessed July 10, 2012
- 5.US Renal Data System: Chronic disease in the general population. In: 2011 USRDS Annual Data Report Available at: http://www.usrds.org/2011/pdf/v1_ch01_11.pdf Accessed June 30, 2012
- 6.US Renal Data System: 2011 USRD Annual Data Report: Atlas of End-Stage Renal Disease in the United States, Vol. 2. Available at: http://www.usrds.org/2011/slides/indiv/v2index.html Accessed June 30, 2012
- 7.Ploth DW, Shepp PH, Counts C, Hutchison F: Prospective analysis of global costs for maintenance of patients with ESRD. Am J Kidney Dis 42: 12–21, 2003 [DOI] [PubMed] [Google Scholar]
- 8.Silverstein MD, Heit JA, Mohr DN, Petterson TM, O’Fallon WM, Melton LJ, 3rd: Trends in the incidence of deep vein thrombosis and pulmonary embolism: A 25-year population-based study. Arch Intern Med 158: 585–593, 1998 [DOI] [PubMed] [Google Scholar]
- 9.Tapson VF: Acute pulmonary embolism. N Engl J Med 358: 1037–1052, 2008 [DOI] [PubMed] [Google Scholar]
- 10.Horlander KT, Mannino DM, Leeper KV: Pulmonary embolism mortality in the United States, 1979-1998: An analysis using multiple-cause mortality data. Arch Intern Med 163: 1711–1717, 2003 [DOI] [PubMed] [Google Scholar]
- 11.Wattanakit K, Cushman M, Stehman-Breen C, Heckbert SR, Folsom AR: Chronic kidney disease increases risk for venous thromboembolism. J Am Soc Nephrol 19: 135–140, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wattanakit K, Cushman M: Chronic kidney disease and venous thromboembolism: Epidemiology and mechanisms. Curr Opin Pulm Med 15: 408–412, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Folsom AR, Lutsey PL, Astor BC, Wattanakit K, Heckbert SR, Cushman M, Atherosclerosis Risk in Communities Study : Chronic kidney disease and venous thromboembolism: A prospective study. Nephrol Dial Transplant 25: 3296–3301, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mahmoodi BK, Gansevoort RT, Veeger NJ, Matthews AG, Navis G, Hillege HL, van der Meer J, Prevention of Renal and Vascular End-stage Disease (PREVEND) Study Group : Microalbuminuria and risk of venous thromboembolism. JAMA 301: 1790–1797, 2009 [DOI] [PubMed] [Google Scholar]
- 15.Tveit DP, Hypolite IO, Hshieh P, Cruess D, Agodoa LY, Welch PG, Abbott KC: Chronic dialysis patients have high risk for pulmonary embolism. Am J Kidney Dis 39: 1011–1017, 2002 [DOI] [PubMed] [Google Scholar]
- 16.Casserly LF, Reddy SM, Dember LM: Venous thromboembolism in end-stage renal disease. Am J Kidney Dis 36: 405–411, 2000 [DOI] [PubMed] [Google Scholar]
- 17.Guntupalli K, Soffer O, Baciewicz P: Pulmonary embolism in end stage renal disease. Intensive Care Med 16: 405–407, 1990 [DOI] [PubMed] [Google Scholar]
- 18.Ifudu O, Delaney VB, Barth RH, Friedman EA: Deep vein thrombosis in end-stage renal disease. ASAIO J 40: 103–105, 1994 [PubMed] [Google Scholar]
- 19.Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project: Introduction to the HCUP Nationwide Inpatient Sample (NIS), 2008. Available at: http://www.hcup-us.ahrq.gov/db/nation/nis/NIS_Introduction_2008.jsp Accessed November 26, 2011
- 20.White RH, Garcia M, Sadeghi B, Tancredi DJ, Zrelak P, Cuny J, Sama P, Gammon H, Schmaltz S, Romano PS: Evaluation of the predictive value of ICD-9-CM coded administrative data for venous thromboembolism in the United States. Thromb Res 126: 61–67, 2010 [DOI] [PubMed] [Google Scholar]
- 21.Deyo RA, Cherkin DC, Ciol MA: Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 45: 613–619, 1992 [DOI] [PubMed] [Google Scholar]
- 22.US Renal Data System: Chronic kidney disease identified in the claims data. In: Atlas of Chronic Kidney Disease Available at: http://www.usrds.org/2009/pdf/V1_03_09.pdf. Accessed November 10, 2011
- 23.US Census Bureau: American FactFinder. Available at: http://factfinder.census.gov Accessed November 10, 2011
- 24.Shlipak MG, Fried LF, Crump C, Bleyer AJ, Manolio TA, Tracy RP, Furberg CD, Psaty BM: Elevations of inflammatory and procoagulant biomarkers in elderly persons with renal insufficiency. Circulation 107: 87–92, 2003 [DOI] [PubMed] [Google Scholar]
- 25.Thijs A, Nanayakkara PW, Ter Wee PM, Huijgens PC, van Guldener C, Stehouwer CD: Mild-to-moderate renal impairment is associated with platelet activation: A cross-sectional study. Clin Nephrol 70: 325–331, 2008 [PubMed] [Google Scholar]
- 26.Hrafnkelsdóttir T, Ottosson P, Gudnason T, Samuelsson O, Jern S: Impaired endothelial release of tissue-type plasminogen activator in patients with chronic kidney disease and hypertension. Hypertension 44: 300–304, 2004 [DOI] [PubMed] [Google Scholar]
- 27.Pawlak K, Tankiewicz J, Mysliwiec M, Pawlak D: Tissue factor/its pathway inhibitor system and kynurenines in chronic kidney disease patients on conservative treatment. Blood Coagul Fibrinolysis 20: 590–594, 2009 [DOI] [PubMed] [Google Scholar]
- 28.Matsuo T, Koide M, Kario K, Suzuki S, Matsuo M: Extrinsic coagulation factors and tissue factor pathway inhibitor in end-stage chronic renal failure. Haemostasis 27: 163–167, 1997 [DOI] [PubMed] [Google Scholar]
- 29.Takagi M, Wada H, Mukai K, Minamikawa K, Wakita Y, Deguchi K, Junji N, Hayashi T, Suzuki K, Shiku H: Increased activated protein C: Protein C inhibitor complex and decreased protein C inhibitor levels in patients with chronic renal failure on maintenance hemodialysis. Clin Appl Thromb Hemost 5: 113–116, 1999 [DOI] [PubMed] [Google Scholar]
- 30.Tomura S, Nakamura Y, Deguchi F, Ando R, Chida Y, Marumo F: Coagulation and fibrinolysis in patients with chronic renal failure undergoing conservative treatment. Thromb Res 64: 81–90, 1991 [DOI] [PubMed] [Google Scholar]
- 31.Gawaz MP, Dobos G, Späth M, Schollmeyer P, Gurland HJ, Mujais SK: Impaired function of platelet membrane glycoprotein IIb-IIIa in end-stage renal disease. J Am Soc Nephrol 5: 36–46, 1994 [DOI] [PubMed] [Google Scholar]
- 32.Pawlak D, Malyszko J, Malyszko JS, Pawlak K, Buczko W, Mysliwiec M: Peripheral serotonergic system in uremia. Thromb Res 83: 189–194, 1996 [DOI] [PubMed] [Google Scholar]
- 33.Di Minno G, Cerbone A, Usberti M, Cianciaruso B, Cortese A, Farace MJ, Martinez J, Murphy S: Platelet dysfunction in uremia. II. Correction by arachidonic acid of the impaired exposure of fibrinogen receptors by adenosine diphosphate or collagen. J Lab Clin Med 108: 246–252, 1986 [PubMed] [Google Scholar]
- 34.Noris M, Benigni A, Boccardo P, Aiello S, Gaspari F, Todeschini M, Figliuzzi M, Remuzzi G: Enhanced nitric oxide synthesis in uremia: Implications for platelet dysfunction and dialysis hypotension. Kidney Int 44: 445–450, 1993 [DOI] [PubMed] [Google Scholar]
- 35.Jalal DI, Chonchol M, Targher G: Disorders of hemostasis associated with chronic kidney disease. Semin Thromb Hemost 36: 34–40, 2010 [DOI] [PubMed] [Google Scholar]
- 36.Bostom AG, Lathrop L: Hyperhomocysteinemia in end-stage renal disease: Prevalence, etiology, and potential relationship to arteriosclerotic outcomes. Kidney Int 52: 10–20, 1997 [DOI] [PubMed] [Google Scholar]
- 37.Sagripanti A, Cupisti A, Baicchi U, Ferdeghini M, Morelli E, Barsotti G: Plasma parameters of the prothrombotic state in chronic uremia. Nephron 63: 273–278, 1993 [DOI] [PubMed] [Google Scholar]
- 38.Camici M, Evangelisti L, Balestri P, Cioni L, Rindi P, Sagripanti A, Meriggioli M, Giordani R: Coagulation activation in extracorporeal hemodialysis. Int J Artif Organs 20: 163–165, 1997 [PubMed] [Google Scholar]
- 39.Churchill DN, Muirhead N, Goldstein M, Posen G, Fay W, Beecroft ML, Gorman J, Taylor DW: Probability of thrombosis of vascular access among hemodialysis patients treated with recombinant human erythropoietin. J Am Soc Nephrol 4: 1809–1813, 1994 [DOI] [PubMed] [Google Scholar]
- 40.Ghisdal L, Broeders N, Wissing KM, Mena JM, Lemy A, Wijns W, Pradier O, Donckier V, Racapé J, Vereerstraeten P, Abramowicz D: Thrombophilic factors in Stage V chronic kidney disease patients are largely corrected by renal transplantation. Nephrol Dial Transplant 26: 2700–2705, 2011 [DOI] [PubMed] [Google Scholar]
- 41.Mossey RT, Kasabian AA, Wilkes BM, Mailloux LU, Susin M, Bluestone PA: Pulmonary embolism: Low incidence in chronic renal failure. Arch Intern Med 142: 1646–1648, 1982 [PubMed] [Google Scholar]
- 42.Wiesholzer M, Kitzwögerer M, Harm F, Barbieri G, Hauser AC, Pribasnig A, Bankl H, Balcke P: Prevalence of preterminal pulmonary thromboembolism among patients on maintenance hemodialysis treatment before and after introduction of recombinant erythropoietin. Am J Kidney Dis 33: 702–708, 1999 [DOI] [PubMed] [Google Scholar]
- 43.Daneschvar HL, Seddighzadeh A, Piazza G, Goldhaber SZ: Deep vein thrombosis in patients with chronic kidney disease. Thromb Haemost 99: 1035–1039, 2008 [DOI] [PubMed] [Google Scholar]
- 44.Monreal M, Falgá C, Valle R, Barba R, Bosco J, Beato JL, Maestre A, RIETE Investigators : Venous thromboembolism in patients with renal insufficiency: Findings from the RIETE Registry. Am J Med 119: 1073–1079, 2006 [DOI] [PubMed] [Google Scholar]