Comparative clinical outcomes after exposure to alternate low osmolar contrast media (LOCM) during invasive coronary angiography (ICA) and/or percutaneous coronary intervention (PCI) have been incompletely examined. From a retrospective multicenter observational study, we identified 107,994 adults without previous hemodialysis undergoing ICA and/or PCI with iohexol, iopamidol, or ioversol. We created a propensity score for contrast media type using age, gender, coverage status, route of hospitalization, illness severity, physician specialty, co-morbidities, and procedure type. Propensity matching was performed in a 1:1 fashion for iohexol (n = 10,204) and iopamidol (n = 10,204) and in a 1:1 fashion for iohexol (n = 19,482) and ioversol (n = 19,482). Groups were examined for differences in in-hospital mortality or subsequent hemodialysis, length of stay, and 30-day readmission for contrast-induced nephropathy (CIN). Compared to patients exposed to iohexol, no differences were observed for patients exposed to iopamidol or ioversol for in-hospital hemodialysis (0.5% vs 0.4%, p = 0.45; 0.3% vs 0.5%, p = 0.05), in-hospital mortality (0.7% vs 0.6%, p = 0.60; 0.5% vs 0.6%, p = 0.42), or composite hemodialysis or mortality (1.1% vs 1.0%, p = 0.58; 0.8% vs 1.0%, p = 0.06); for hospital length of stay (2.9 ± 2.7 vs 2.9 ± 2.7 days, p = 0.05; 2.8 ± 2.6 vs 2.9 ± 3.1 days, p = 0.35); or for 30-day readmission for CIN (0.1% vs 0.1%, p = 0.82; 0.1% vs 0.1%, p = 0.52). In conclusion, for patients undergoing ICA and/or PCI exposed to alternate LOCM, in-hospital death, need for hemodialysis, or readmission for CIN are uncommon, with no apparent clinical advantage among LOCM agents.
The use of iodinated contrast media in patients undergoing invasive coronary angiography (ICA) and/or percutaneous coronary intervention (PCI) may result in contrast-induced nephropathy (CIN), which is associated with significantly increased morbidity and mortality. Current societal guidelines recommend the use of other low osmolar contrast media (LOCM) over iohexol in patients with chronic renal insufficiency because the latter has been associated with potentially higher rates of acute kidney injury compared to iso-osmolar contrast media in previous meta-analyses. However, to date no large-scale direct comparisons for different LOCM types have been performed in patients undergoing ICA and/or PCI. To this end, we examined within a retrospective multicenter registry of 107,994 patients undergoing ICA and/or PCI the comparative clinical outcomes of matched patients exposed to iohexol, iopamidol, and ioversol.
Methods
We performed a retrospective analysis of hospital data for 107,994 patients in the Premier Perspective Database (Premier, Charlotte, North Carolina). This voluntary fee-supported database represents approximately 5.5 million patient discharges per year from >600 geographically disperse hospitals from all regions of the United States. For each patient, the database contains a date-stamped log of all billed items including procedures, medications and laboratory tests. Primary and secondary diagnosis and procedural codes are collected by the International Classification of Diseases, Ninth Revision, Clinical Modification . Identifier-linked enrollment files provided demographic and payer information. All patient records were made anonymous in compliance with the Health Insurance Portability and Accountability Act. All sites were in compliance with institutional review board requirements.
Patients were identified for inclusion in this analysis from records of all hospital discharges occurring from January 1, 2007 through December 31, 2008, with ICA with or without PCI deemed as the index procedure. To be eligible for inclusion, patients were (1) ≥18 years of age; (2) hospitalized; (3) undergoing diagnostic ICA and/or PCI procedures during hospitalization; and (4) exposed to 1 of 3 different contrast media: iohexol, ioversol, or iopamidol. Patients were excluded if they (1) underwent hemodialysis before the index procedure or (2) were exposed to multiple different contrast media during the index hospitalization. Patients could be included if they had undergone >1 procedure with the same contrast media during the hospitalization, if the first procedure using contrast media was the index ICA and/or PCI procedure. Eligible patients were categorized based on the contrast medium used for the procedure during index hospitalization.
For each patient, we collected the following information: age, gender, payer status, route of hospitalization, diagnosis-related group illness severity, attending physician specialty, presence of co-morbidities, and type of procedure performed. For payer status, particular focus was paid to patients covered by Medicare and Medicaid. Route of hospitalization was defined as admission to the hospital from a nonhealth care facility, transfer from an outside hospital, or through the emergency room. All patient-refined diagnosis-related groups enabled determination of illness severity, which was defined as minor, moderate, major, or extreme. Attending physician specialty was categorized as cardiologist versus noncardiologist. Co-morbidities included diabetes, chronic renal insufficiency, acute renal insufficiency, congestive heart failure, myocardial infarction, nephritis or nephrosis, hypertension, angina pectoris, anemia, or cancer. Presence of co-morbidities was based on physician diagnoses listed in the medical records. Procedure type consisted of diagnostic ICA and/or PCI. Additional procedures were identified by procedural codes based on temporal use of contrast media.
Primary outcome measurements included rates of in-hospital mortality, in-hospital hemodialysis, hospital length of stay (LOS), and readmission for CIN within 30 days. Inpatient hemodialysis was defined as any hemodialysis during hospital admission after the index procedure. LOS for the initial hospitalization was defined as days from hospital admission to discharge. LOS in the intensive care unit during the initial hospitalization was also measured. Readmission for CIN was defined as a subsequent hospitalization within 30 days of discharge for which the primary or secondary diagnoses included new renal insufficiency ( International Classification of Diseases, Ninth Revision, Clinical Modification 585 or 586), need for new dialysis ( International Classification of Diseases, Ninth Revision, Clinical Modification V56A or V45B), or procedure for new dialysis (procedure code v9211, v9212, v9200, v9531, or v9532).
Baseline demographics and clinical characteristics were described with standard summary statistics by means and proportions. Covariates were compared using chi-square tests for categorical variables and t tests and Mann–Whitney tests for continuous variables. Wilcoxon signed-rank test was used to assess differences in LOS, with the McNemar test used to assess differences in hemodialysis, mortality, and readmission rates.
Propensity scores were created for contrast media type for each patient included in the analysis based on a comprehensive logistic regression model that predicted the probability of being exposed to iohexol at the time of ICA with or without PCI. Propensity score models included age, gender payer type, point of origin, all patient-refined diagnosis-related group severity of illness, attending physician specialty, co-morbidities, and type of procedure. Patients exposed to iohexol were matched to patients receiving ioversol or iopamidol on propensity score using the nearest available pair method, with the balancing property achieved to a minimum of 3 decimal places. We determined the success of the propensity score to decrease bias by evaluating standardized differences in covariates between cohorts after matching, with standardized differences <10% considered acceptable and indicative of a successful match. Continuous variables were summarized employing mean ± SD and were compared by paired t test or Wilcoxon signed-ranks test adjusting for the matched pair. Categorical variables were summarized by frequencies and percentages and compared using the Cochran-Mantel-Haenszel test that adjusted for the matched pair. All statistical analyses tested a 2-sided hypothesis of no difference between treatment cohorts at a significance level of 0.05. Analyses were performed using SAS 9.2 (SAS Institute, Cary, North Carolina).
Results
Baseline patient characteristics of the 107,994 patients comprising the study population are listed in Table 1 , stratified by those exposed to iohexol (n = 20,136), iopamidol (n = 21,539) or ioversol (n = 66,319) at the time of ICA and/or PCI. Significant differences existed between groups for age, gender, payer, point of origin, illness severity, attending physician specialty, co-morbidities, and type of index procedure. Table 2 presents comparisons of in-hospital outcomes, LOSs, and rates of 30-day readmission for CIN, which differed significantly among groups.
| Variable | Iohexol | Iopamidol | Ioversol | p Value |
|---|---|---|---|---|
| (n = 20,136) | (n = 21,539) | (n = 66,31) | ||
| Demographics | ||||
| Age ≥65 years | 43.6% | 44.6% | 43.4% | 0.009 |
| Men | 57.8% | 58.8% | 58.9% | 0.03 |
| Primary payer | ||||
| Medicare | 47.6% | 50.5% | 47.6% | <0.001 |
| Medicaid | 4.8% | 5.3% | 6.0% | <0.001 |
| Commercial | 36.3% | 32.9% | 35.8% | <0.001 |
| Self-pay/indigent | 7.3% | 8.5% | 6.9% | <0.001 |
| Any other payer | 4.0% | 2.9% | 3.7% | <0.001 |
| Point of origin | ||||
| Nonhealth care facility | 32.9% | 24.0% | 29.3% | <0.001 |
| Outside transfer | 11.2% | 18.8% | 14.6% | <0.001 |
| Emergency room | 52.4% | 51.6% | 47.7% | <0.001 |
| Other/unknown | 3.5% | 5.7% | 8.4% | <0.001 |
| All patient-refined diagnosis-related group severity of illness | ||||
| Minor (1) | 42.8% | 38.6% | 44.1% | <0.001 |
| Moderate (2) | 38.3% | 39.0% | 37.7% | 0.003 |
| Major (3) | 16.3% | 18.8% | 15.5% | <0.001 |
| Extreme (4) | 2.6% | 3.6% | 2.7% | <0.001 |
| Attending physician specialty | ||||
| Cardiology | 65.5% | 57.5% | 65.8% | <0.001 |
| Noncardiology | 34.5% | 42.5% | 34.2% | <0.001 |
| Co-morbidities (before study) | ||||
| Diabetes mellitus | 30.3% | 33.5% | 31.2% | <0.001 |
| Chronic renal insufficiency | 2.6% | 4.7% | 2.9% | <0.001 |
| Acute renal insufficiency | 1.3% | 2.2% | 2.0% | <0.001 |
| Heart failure | 14.0% | 15.9% | 14.4% | <0.001 |
| Myocardial infarction | 27.3% | 31.5% | 30.8% | <0.001 |
| Nephritis or nephrosis | 5.7% | 9.3% | 5.9% | <0.001 |
| Hypertension | 61.9% | 63.8% | 64.2% | <0.001 |
| Angina pectoris | 8.7% | 7.2% | 10.3% | <0.001 |
| Anemia | 8.4% | 10.2% | 7.7% | <0.001 |
| Cancer | 1.4% | 2.0% | 1.6% | <0.001 |
| Type of procedure | ||||
| Diagnostic coronary angiography | 45.9% | 48.4% | 44.1% | <0.001 |
| Percutaneous coronary intervention | 54.1% | 51.6% | 55.9% | <0.001 |
| Variable | Iohexol | Iopamidol | Ioversol | p Value |
|---|---|---|---|---|
| (n = 20,136) | (n = 21,539) | (n = 66,319) | ||
| In-hospital events | ||||
| Hemodialysis | 0.34% | 0.48% | 0.49% | 0.02 |
| Mortality | 0.6% | 0.7% | 0.7% | 0.17 |
| Total events | 0.9% | 1.1% | 1.1% | 0.005 |
| Length of stay | ||||
| Intensive care unit stay (days) | 2.0 ± 2.2 | 2.1 ± 2.3 | 2.1 ± 2.2 | 0.06 |
| Total stay (days) | 2.8 ± 2.6 | 3.0 ± 3.0 | 2.9 ± 3.0 | <0.001 |
| Contrast-induced nephropathy 30-day readmission | 0.1% | 0.1% | 0.1% | 0.77 |
Table 3 lists characteristics of matched patients exposed to iohexol (n = 19,482) versus ioversol (n = 19,482). Despite matching, small differences were noted for patients admitted to the hospital from an “other/unknown” location who were exposed to iohexol or ioversol, whereas small differences between iohexol- and ioversol-exposed patients also existed in rates of diabetes mellitus and previous myocardial infarction. Table 4 lists characteristics of matched patients exposed to iohexol (n = 10,204) versus iopamidol (n = 10,204). Despite matching, slight differences were observed for patients whose payer was listed as “self-pay/indigent” or “any other payer” and in rates of previous myocardial infarction.
| Variable | Iohexol | Ioversol | Standardized Difference | p Value |
|---|---|---|---|---|
| (n = 19,482) | (n = 19,482) | |||
| Demographics | ||||
| Age ≥65 years | 43.7% | 44.2% | 1.1% | 0.22 |
| Men | 58.0% | 58.2% | 0.5% | 0.53 |
| Primary payer | ||||
| Medicare | 47.6% | 47.4% | 0.4% | 0.67 |
| Medicaid | 4.9% | 4.9% | 0.0% | 0.98 |
| Commercial | 36.3% | 37.0% | 1.4% | 0.09 |
| Self-pay/indigent | 7.2% | 7.1% | 0.5% | 0.56 |
| Any other payer | 4.0% | 3.7% | 1.8% | 0.06 |
| Point of origin | ||||
| Nonhealth care facility | 32.8% | 32.9% | 0.2% | 0.77 |
| Outside transfer | 11.5% | 11.9% | 1.3% | 0.12 |
| Emergency room | 52.1% | 52.1% | 0.1% | 0.95 |
| Other/unknown | 3.6% | 3.1% | 3.1% | <0.001 |
| All patient-refined diagnosis-related group severity of illness | ||||
| Minor (1) | 43.0% | 42.8% | 0.5% | 0.55 |
| Moderate (2) | 38.4% | 39.2% | 1.6% | 0.06 |
| Major (3) | 16.1% | 15.6% | 1.3% | 0.18 |
| Extreme (4) | 2.6% | 2.5% | 0.6% | 0.53 |
| Attending physician specialty | ||||
| Cardiology | 65.5% | 66.2% | 1.4% | 0.09 |
| Noncardiology | 34.5% | 33.8% | 1.4% | 0.09 |
| Co-morbidities (before procedure) | ||||
| Diabetes mellitus | 30.2% | 29.4% | 1.9% | 0.03 |
| Chronic renal insufficiency | 2.6% | 2.5% | 0.8% | 0.39 |
| Acute renal insufficiency | 1.4% | 1.3% | 1.0% | 0.34 |
| Heart failure | 13.9% | 13.3% | 1.6% | 0.07 |
| Myocardial infarction | 27.5% | 28.3% | 1.9% | 0.02 |
| Nephritis or nephrosis | 5.7% | 5.4% | 1.0% | 0.30 |
| Hypertension | 62.2% | 61.4% | 1.7% | 0.047 |
| Angina pectoris | 8.8% | 8.4% | 1.5% | 0.28 |
| Anemia | 8.1% | 7.9% | 0.8% | 0.41 |
| Cancer | 1.4% | 1.4% | 0.3% | 0.76 |
| Type of procedure | 1.0 | |||
| Diagnostic coronary angiography | 46.2% | 46.2% | <0.1% | — |
| Percutaneous coronary intervention | 53.8% | 53.8% | <0.1% | — |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
