Acute renal embolus is rarely reported in the medical literature. Two series have been reported recently. Domanovits et al3 described the clinical characteristics of 17 patients with renal infarction over a 45-month period, 11 (65%) of whom had atrial fibrillation and were thus assumed to have an embolic event. Korzets et al8 presented 11 cases of renal infarction in 10 patients, over a 36-month period. Five of these patients had atrial fibrillation. In 1978, Lessman et al9 reported 17 patients during 14 years with emboli to major renal arteries based on clinical data from 14 patients and autopsy results from 3. In a postmortem study, Hoxie and Coggin7 reported an incidence of 1.4%; most of the cases were not diagnosed antemortem, which suggests that the condition is grossly underdiagnosed. That study was published in 1940, when current imaging techniques were not available. Thus we examined the number of cases of renal embolus diagnosed in the antemortem clinical records of 2 medical centers in Israel, from 1984 to 2002, to learn more about the clinical characteristics of these patients and to assist in the early diagnosis of this potentially treatable condition.
We examined the medical records of all patients with a discharge diagnosis of renal infarction and atrial fibrillation who were treated in Kaplan Medical Center in Rehovot, Israel, and Sheba Medical Center in Tel Aviv, Israel, between 1984 and 2002. The diagnosis of renal embolus was based on the presence of atrial fibrillation either on admission or during the hospitalization; a suggestive clinical picture of severe loin or abdominal pain of no other cause; the presence of microscopic hematuria, fever, nausea, oliguria, or a combination of these symptoms and signs together with radiologic evidence of an embolus based on a renal scan, ultrasound, or computed tomography (CT) scan of the kidneys or renal angiography; and an elevated level of lactate dehydrogenase (LDH). Evidence for a renal embolus on isotope scan was the presence of a perfusion deficit and a nonobstructed, nonfunctioning kidney. Ultrasound evidence for a renal embolus consisted of hypoechogenic and hyperechogenic areas in the renal cortex with perirenal fluid. On CT scan of a kidney with a renal embolus, there was a wedge-shaped zone of peripheral diminished density without enhancement by intravenous contrast medium. Angiography was considered positive if there was demonstration of an occlusion or filling defect in the renal artery. To be included in the study we also required that the renal emboli be 1 of the diagnoses on the discharge summary. Laboratory investigations were not part of the inclusion criteria. The data collected included the patient's age; sex; previous medical history; current medical treatment; time from admission to diagnosis; method of diagnosis; international normalized ratio (INR) at the time of admission; laboratory parameters including complete blood count, renal function and serum electrolytes, liver enzymes, creatinine phosphokinase; the electrocardiogram on admission; and changes during the course of the hospitalization. On urinalysis microscopic hematuria was defined as the presence of at least 3 red blood cells per high-power field examination of the urine. Details of the treatment administered and outcome were noted.
Statistical analysis was performed with the SPSS program version 9. The Student t-test was performed for numerical variables and the χ-squared test for nonparametric variables. Values are stated ± standard deviation.
Forty-four cases of renal embolus were identified from a retrospective review of patient files (Table 1). There were 23 female patients and 21 males. The average age of the patients was 69.5 ± 12.6 years. The duration of symptoms before presentation was less than 24 hours in 8 cases, from 24 to 48 hours in 17 cases, 48-72 hours in 6 cases, and more than 72 hours in 8 cases. (No data were available on duration of symptoms before presentation in 5 cases.) Seven (16%) of the patients had a history of a previous embolic event. Thirty (68%) patients had generalized abdominal pain, 14 (32%) had lumbar pain, 3 (7%) had right upper quadrant pain, 18 (41%) had fever, 3 (7%) had both fever and abdominal pain, and 6 (14%) had both lumbar pain and fever. Nineteen (43%) patients experienced vomiting, 18 (41%) had fever, and only 3 (7%) had oliguria. The reason for admission was suspected renal embolism in 17 patients, but in another 7 the initial diagnosis was a mesenteric ischemic event.
Forty patients had a previous history of atrial fibrillation. Four patients had a new onset of rapid atrial fibrillation, which was the reason for admission. One of the patients had a new 2-week history of atrial fibrillation following coronary artery bypass surgery. Twenty patients had chronic atrial fibrillation; no data are available on the duration of the fibrillation. Four patients had a history of paroxysmal atrial fibrillation, although no data were available on the admission electrocardiogram (ECG). Seven patients with a history of paroxysmal atrial fibrillation were in sinus rhythm on admission. In 11 patients with paroxysmal atrial fibrillation, atrial fibrillation was noted on the admission ECG, although the duration of this finding was unclear. Thus the majority of patients in the current study had atrial fibrillation either on admission or during the hospitalization.
The left kidney was involved in 18 (41%) cases; 4 patients had, in addition, a splenic infarction. Seven patients had a history of previous embolic events, 3 of whom had multiple embolic events, for a total of 10 previous thromboembolic events. Nine patients were being treated with warfarin on admission. One had an INR of 2.2 and another an INR of 2.1 on admission. Six patients had an INR <1.8. Thirty-three patients had atrial fibrillation on the admission ECG and 7 had sinus rhythm; there were no data on the other 4 patients, although they had a history of paroxysmal atrial fibrillation. Oliguria was present in 3 (7%) of 44 patients.
Urinalysis on admission revealed hematuria in 21/39 (54%) and proteinuria in 17/38 (45%). Of the 21 patients with hematuria on admission, 10 had normal renal function, whereas of the 18 patients without hematuria only 5 had normal renal function (p = 0.059, Pearson chi-squared test). The serum urea on admission was 52.71 ± 34.5 mg/dL and increased to 70.8 ± 61.72 mg/dL (p = 0.011). The serum creatinine on admission was 1.79 ± 1.66 mg/dL and increased to 2.43 ± 3.37 mg/dL (p = 0.123).
Data on the serum urea on discharge were available for 38 patients. In 23 patients, the diagnosis of renal embolus was made on admission or the first hospital day; they had a mean serum urea at discharge of 59.9 ± 50.2 mg/dL. In 6 patients the diagnosis was made on day 2; they had a mean serum urea at discharge of 73 ± 58.3 mg/dL. The diagnosis was made on day 3 in 5 patients; they had a mean serum urea of 89.6 ± 107.2 mg/dL. There was no significant difference in the serum urea on discharge between those patients whose diagnosis was made on the first hospital day and those whose diagnosis was made on later days. In addition, there was no significant difference in serum creatinine (data not shown).
The serum LDH level on admission was 1100.1 ± 984.6 U/dL and subsequently 908.1 ± 537.3. This difference was not significant (p = 0.193). Forty-one patients (93%) had a serum LDH greater than 400 U/dL. In 20 patients the LDH was checked on the day of admission. In 5 of these the LDH was normal and it increased the next day-these patients all had the initial pain on the day of admission. The other 15 patients had an elevated LDH on the day of admission, and 9 of these were admitted on the first day of the occurrence of the pain. In contrast, on the second hospital day the LDH was elevated in 15 patients and normal in 1. This suggests that LDH is elevated as a result of the initial infarction, similar to the classic picture seen for myocardial infarction.
The diagnosis was made on admission in 17 (40%) of 42 patients with data available, after 1 day in 6 (14%) patients, after 2 days in 6 (14%) patients, after 3 days in 5 (12%) patients, after 4 days in 3 (7%) patients, and as late as 12 days after admission in 1 patient. The initial diagnosis was renal colic in 9 (20%) patients, a mesenteric event in 7 (16%) patients, pyelonephritis in 4 (9%) patients, nonspecific abdominal pain in 4 (9%) patients, urinary tract infection in 3 (7%) patients, and biliary tract disease in 2 (5%) patients.
Only 22 (50%) patients had an echocardiographic evaluation. On transthoracic echocardiography, 11 patients had no detectable thrombus, 5 had wall-motion abnormality or a decreased left ventricular ejection fraction, 2 had severe mitral stenosis, 1 had left ventricular thrombus, 1 had minimal mitral regurgitation, 1 had mild-moderate tricuspid regurgitation, 1 had severe tricuspid regurgitation, and 1 had a sigmoid interventricular septum. Transesophageal echocardiography was performed in 4 patients. No thrombus was detected in 3 and 1 had severe mitral stenosis.
Thirty-eight patients received treatment with heparin and warfarin, 3 with aspirin alone, and 3 did not receive anticoagulation at all, 1 of whom had a peptic ulcer. In addition 7 patients received thrombolytic therapy with intraarterial streptokinase or urokinase, and 1 patient had a renal angioplasty.
The outcome of the renal embolic event was good for most patients, as shown in Table 2. Twenty-three (61%) of 38 patients for whom data were available had normal renal function on follow-up investigations with a mean time of 161 ± 223 weeks (range, 0.4-780 wk). Five (13%) of the 38 patients had mild renal impairment (creatinine from 1.5 to 2.0 mg/dL), and 7 (18%) of the 38 patients had serum creatinine >2 mg/dL. Five (11%) of the 44 total patients died during the following month. Three (8%) had dialysis-dependent acute renal failure. The 3 patients with end-stage renal failure died. The mean creatinine was 5.11 ± 6.37 mg/dL in those patients who died versus 1.66 ± 1.02 mg/dL in those who survived (p = 0.043). This implies that renal failure is a predictor of severity.
There was a tendency for repeat embolic events. Seven patients had a previous history of embolism and 11 patients developed a subsequent thromboembolic event (data not shown), including 3 patients who had ischemic cerebrovascular events (within 3 days to 3 weeks of the index renal embolus). At least 1 of these 3 patients was being treated with warfarin (INR 1.2) at the time of the repeat embolic event.
The most common diagnostic technique was a renal isotope scan, which was abnormal in 36 (97%) of 37 cases. Contrast-enhanced CT scan was diagnostic in 12 (80%) of 15 cases, whereas ultrasound was positive in only 3 (11%) of 27 cases. Angiography was positive in 10 (100%) of 10 cases.
Combining our group of patients with the cases in the literature provides data on a total of 89 patients. The combined data are shown in Tables 3, 4, and 5. The 89 patients have a mean age of 65.7 years. The gender distribution was nearly equal with 47% female patients. In addition, kidney involvement was nearly equal. The left kidney alone was involved in 38 of 89 (43%) cases, and the right kidney alone in 40 of 89 (45%) cases. Two additional cases had bilateral involvement of the kidneys. Of the clinical symptoms, pain was present in 81 of 89 (91%) cases, vomiting in 40%, fever in 49%, and oliguria in 16%. Pooling of the laboratory investigations revealed hematuria in 72%, leukocytosis in 85%, elevated LDH in 91%, and increased serum creatinine in 53%. The accuracy of the various diagnostic techniques is 96% for a renal isotope scan, 88% for a CT scan, and 100% for angiography. The diagnosis was made on admission in only 33% of the cases, and was made 48 hours after admission in 42 of 87 (48%) cases. Renal function was preserved in 61% of patients. Dialysis-dependent acute renal failure occurred in 9.1%, and the combined mortality was 10.2%. Of the 2 cases with bilateral involvement, 1 recovered and the other died of sepsis and required dialysis.
We present here our experience of 44 cases of acute renal embolus in 2 medical centers in central Israel over 18 years. The diagnosis is based on the clinical picture together with radiologic evidence of an embolus (renal scan, ultrasound, CT, or angiography) occurring in the presence of atrial fibrillation, either on admission or during the hospitalization. These cases are probably the result of atrial fibrillation and thrombus formation either on admission or during the index hospitalization.
The fact that our study is retrospective requires us to accept the discharge diagnosis of our medical centers as accurate. Furthermore, the gold standard investigation of renal artery angiography was performed in only 10 cases. We believe, however, that the diagnosis of renal infarction is likely and does reflect common medical practice. Renal angiography is not performed on a routine basis, and the improvements in CT technology (especially with the use of spiral CT) in recent years have resulted in a marked improvement in our diagnostic capability. Indeed, our experience with diagnostic tests for renal emboli is similar to that of other cases in the literature8,9. Renal angiography was performed in only those cases that presented in the 1980s, which underlines the acceptance of other diagnostic techniques into clinical practice.
Most cases presented with the patient in the sixth decade of life and commonly presented as abdominal or flank pain, fever, leukocytosis, and hematuria. This constellation of complaints can cause diagnostic confusion; initial diagnoses included renal colic, pyelonephritis, abdominal pain of uncertain cause, an acute mesenteric event, urinary tract infection, and biliary tract disease. The diagnosis of renal embolus was made on admission in only 17 of 42 patients (40%). Most of the diagnoses were made within the first 3 days following admission; only 9 were made later than 3 days after admission. Thus, in clinical practice renal embolism is often not diagnosed on admission. We suggest that an increased awareness of this not-so-rare entity will enable the diagnosis to be made more readily.
Laboratory investigations performed included urinalysis and standard blood investigations. Since we include data from 2 different medical centers over 2 decades, it is not possible to include the normal values for the blood investigations in this report, but an approximation can be made by comparison with published normal values. In the current study microscopic hematuria was present in 21 of 39 cases (54%). Microscopic hematuria was present in 25 of 35 cases (71%) in the previously reported literature, which is consistent with our finding that a significant proportion of patients may not have significant hematuria.
Grouping together our results with the other cases in the literature, microscopic hematuria was present in 59 of 82 cases (72%). The absence of hematuria may be due to the decrease in renal blood flow to the infarcted area, causing a decrease in glomerular filtration and urine flow. Indeed, there was almost a significant difference in the patients with hematuria who had normal renal function (10 of 21) compared to the patients without hematuria who had normal renal function (5 of 18, p = 0.059), indicating that the absence of hematuria may suggest a more serious loss of renal function. The serum urea increased significantly from admission to the next day, but there was no significant change in the serum creatinine. The LDH was elevated but did not change significantly from the first to the second day of admission. Although Lessman et al9 found an elevated LDH in all 17 of their patients, in the current series the patients with normal LDH on admission tended to be those who presented earlier. The findings of Lessman et al may reflect the fact that the diagnosis was made later; indeed, nearly 25% of their patients experienced no pain, which may have contributed to the late presentation. In the current series, 23/42 (55%) patients had the diagnosis made either on admission or the first hospital day. It has been suggested that measurement of urinary LDH is of value10. This enzyme is too large to be filtered by the kidney, and its excretion rate is normal in extrarenal disorders. In contrast, in renal infarction and renal transplant rejection there is an increase in the excretion rate from direct renal origin. We do not have any data on urinary LDH.
The diagnosis can be made by use of renal arteriography, ultrasound, CT scan of the abdomen, a renal isotope scan, or excretory urography. Diagnosis of renal artery embolism is established by the demonstration of an occlusion or filling defect in the renal artery on angiography. Renal angiography was positive in all 10 cases performed in our study. This may be assumed to be the gold standard for diagnosis, similar to the situation for the diagnosis of a pulmonary embolism, but of course it has the disadvantage that it is an invasive procedure. Our current clinical practice does not include the use of renal angiography.
Ultrasound seems to be unreliable for diagnosing this condition. In our experience, of the 27 patients who underwent an ultrasound of the abdomen urgently, the diagnosis of renal infarction was made in only 3 (11%), and only after performing other confirmatory investigations, such as CT and a renal isotope scan. In the remainder the ultrasound was reported as normal. Ultrasound was also normal in a case report of a 71-year-old woman2. No data were provided on ultrasound as a diagnostic tool for renal embolus or infarction in the 3 series noted previously3,8,9. A 2001 report suggests that power Doppler sonography may be useful in the diagnosis of renal infarction15. Although ultrasound has the advantage of being inexpensive and widely available, we believe that it should not be used for the diagnosis of renal embolism due to the low sensitivity, and that any abnormal findings need to be confirmed by additional investigations.
The diagnosis of a renal embolism is suggested by the demonstration of a nonobstructed, nonfunctioning kidney. This may be demonstrated by excretion urography (IVP) or a nuclear renal scan. Excretory urography shows a normal or enlarged kidney with no or poor opacification. A nephrogram phase is typically absent, even on delayed films, except in situations of severe or long-standing renal obstruction, renal vein thrombosis, chronic renal failure, renal transplant rejection, or acute pyelonephritis4. A nonopacified kidney of normal or increased size with a healthy pelvicaliceal system revealed by retrograde pyelography is considered diagnostic of renal artery occlusion caused by embolus or thrombus13. Excretion urography was not performed in our study but was positive in 13 of 13 cases in a previous study9. Since the use of helical CT scans, there has been a decrease in the use of excretion urography.
CT scan of the kidney with intravenous contrast media is fast becoming the diagnostic technique of choice for renal embolism. The classic finding is of a wedge-shaped zone of peripheral diminished density without enhancement. The findings on CT have been summarized in a series of 37 cases12. The most common finding was a hypoattenuated area with an associated mass effect, which was present in 32% of cases, followed by the cortical rim sign in 19%. The cortical rim nephrogram sign represents opacification of a rim of functioning nephrons, supplied by capsular collaterals, surrounding an otherwise nonfunctioning kidney. This sign may also be present on high-dose nephrotomography. It may be present in long-standing hydronephrosis, renal vein obstruction, and acute renal failure6. The diagnosis was judged to be clear-cut in 15 of the 37 cases (41%); in the remainder the image was more complex. In addition, a "flip-flop" enhancement pattern (eventual enhancement of initially hypodense infarcts) due to extravasation of contrast from ischemic destruction of the glomerular membrane in delayed contrast-enhanced CT images has also been described12. Two of the 37 cases (5%) were incorrectly diagnosed as tumors.
CT scan of the kidney was a more useful investigation in our experience. Of the 15 tests performed, 12 were positive and 3 negative. In the report of Korzets et al8, 11 cases were presented of renal infarction diagnosed by CT with contrast injection, although there was an initial misdiagnosis in 3 of the cases. A typical CT scan is shown in Figure 1.
Renal scintigraphy is a noninvasive, simple, safe, and available initial study in the detection and diagnosis of renal artery embolism. It offers, in addition, the advantage of assessing the renal function at presentation and comparing the subsequent picture on follow-up. An isotope such as Tc-99m glucoheptonate is often used. This is rapidly cleared from the blood through both glomerular and renal tubular mechanisms. Up to 15% of the administered dose is retained in the kidney. The radiotracer is probably bound to the proximal convoluted tubules and a vascular, excretory, and cortical phase can be assessed with both immediate and delayed images. The use of SPECT evaluation can improve contrast resolution.
Renal scans were positive in 36 of 37 (97%) cases performed in our study. In Lessman and colleagues' series of patients from 19789, 10 of 10 renal isotope scans were positive, which is consistent with our results. We have no information about the false-negative rate and thus cannot comment on the specificity of this test. The imaging techniques described above can demonstrate the presence of an area of decreased perfusion but not the etiology of the infarction. The specific etiology is apparent from the history, physical examination, and laboratory investigations. Other etiologies apart from atrial fibrillation include cholesterol emboli, cocaine use, or aspergillus emboli.
Unenhanced helical CT scan is now thought to be the investigation of choice for the diagnosis of renal colic11, since it can be rapidly performed and can detect nearly all types of renal calculi. In addition it may detect extrarenal causes of abdominal pain including appendicitis, diverticulitis, biliary tract disease, leaking aortic aneurysm, and gynecologic disease. However, it cannot easily detect renal artery thromboembolism. Since the clinical picture of renal artery embolism is similar to that of renal colic (flank pain and microscopic hematuria), the widespread use of unenhanced CT scan needs to be reassessed. We suggest that in those patients with clinical characteristics suggesting renal embolus, such as atrial fibrillation without any or without appropriate anticoagulation, unenhanced CT scans of the abdomen should be followed by enhanced scans if no calculi are found.
The treatment of choice for renal emboli is unclear. In our experience, thrombolysis is not successful. Of the 7 patients who were treated, only 4 had normal renal function on follow-up (1 of whom had a nephrectomy of the affected kidney), 2 died (after developing dialysis-dependent acute renal failure), 1 developed end-stage renal failure, and 1 had mild renal failure. Most of our patients had normal renal function on follow-up. We cannot determine if this is due to damage to the kidney with compensatory hypertrophy or reflects minor damage due to the renal embolus, or is secondary to contrast-induced nephropathy. However, since an acceptable outcome in terms of renal function is achieved in the majority of the patients with treatment consisting of anticoagulation with heparin and subsequently warfarin, we feel it is unjustified to give thrombolytic therapy due to the associated risks of hemorrhage and the poor outcome. Thrombolytic therapy was, however, given in the earlier years of the study. Five of the 7 patients treated with thrombolysis received the treatment between 1984 and 1989 and the other 2 in 1998 and 1999. It may be that the patients treated in the 1980s were diagnosed later, or that thrombolytic therapy was only given to the more severe cases and this could adversely affect the results.
Pooling our results with the others in the literature does not change significantly the major clinical, laboratory, and diagnostic findings, as well as the results of treatment, as shown in Tables 3, 4, and 5. The typical patient is in the sixth decade of life, may have a history of a previous embolic event, and will most likely have pain in the abdomen or loin and microscopic hematuria. In addition, the patient is likely to have leukocytosis and an elevated LDH. Despite the late diagnosis in nearly half the cases, renal function is likely to be preserved. An earlier diagnosis may result in more patients having preserved renal function, but this has not been shown by our data.
This study was not intended to determine the incidence of renal embolus in our population of patients, but to characterize the clinical and laboratory features of this entity. We limited our search to patients with both atrial fibrillation and renal infarction, assuming that this would identify patients with renal embolus. Other groups of patients who are at risk for developing an embolic event, such as those with dilated cardiomyopathy or with atheromatous plaques on the thoracic aorta, have not been included in this study.
Only a minority (9 of 44) patients were treated with warfarin before admission, and only 2 of these had an INR >1.8. Thus the vast majority of patients were either not treated with warfarin or were undercoagulated. It is possible that renal infarction is a common initial presentation of atrial fibrillation, although the relative rarity of this condition in the literature makes this unlikely. We suggest that the high proportion of patients who did not receive anticoagulation underlines the importance of appropriate treatment to prevent embolization in such patients. Nine of the 43 patients had a total of 10 recurrent embolic events (Table 6), suggesting that patients with 1 embolic event are prone to subsequent embolic events. Thus it is important to continue administering warfarin, although we cannot comment on the optimal INR.
In this partial selection of patients from 2 medical centers in central Israel over 18 years, we have identified more cases of renal embolism than the total number of cases published in the world literature. This underlines the fact that this entity is underdiagnosed. The kidney has a rich blood supply, averaging 20% of the cardiac output. In terms of blood flow per 100 g of tissue weight, the blood supply to the kidney is 4 times that of the liver or exercising muscle and 8 times that of coronary blood flow1. Thus it is to be expected that the kidney would be a common site for the occurrence of emboli. However, there are few reports, and population-based studies have not found many cases. In a review5 of a Danish national database between 1980 and 1983, 29,862 patients were discharged with a diagnosis of atrial fibrillation. There were 623 patients with peripheral arterial thromboembolism; 2% of patients had an embolus to the renal artery and 29% to the mesenteric artery. This report was based on data accumulated from discharge diagnoses, and we suspect that it also reflects the underdiagnosis of renal embolus that we have noted.
In summary, we present here 44 cases of renal embolus seen at 2 medical centers in central Israel over 18 years. This number of cases is more than the total cases in the world literature, which indicates that this entity is underdiagnosed. We suggest that a multicenter study is necessary to determine the incidence, the optimal method of diagnosis, and the treatments of choice for this treatable condition. Our current approach to the diagnosis and treatment of renal embolus is shown in Figure 2.