Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Linear Regression”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 955 records · Page 53Linked to original sources
PubMed2005

Correlation between capillary blood flow of retina estimated by SLDF and circulatory parameters of retrobulbar blood vessels estimated by CDI in diabetic patients.

BACKGROUND: The aim of this study was to investigate the correlation between the recordings of scanning laser Doppler flowmetry (SLDF) of the retina and the recordings of color Doppler imaging (CDI) of the retrobulbar circulatory parameters in diabetic patients without diabetic retinopathy. METHODS: Twenty-three diabetic patients without diabetic retinopathy were evaluated using SLDF for the apparent retinal circulation and using CDI for the apparent retrobulbar circulation in the central retinal artery, the central retinal vein and the short posterior ciliary artery. The circulatory parameters estimated in the retinal tissue using SLDF were velocity, volume and flow. The Circulatory parameters that were recorded using the CDI method were peak systolic velocity (PSV), end-diastolic velocity (EDV), mean velocity (MV), pulsatility (PI) and resistivity index (RI). We obtained the correlation coefficients between parameters of SLDF and CDI. Multiple regression analysis was performed with "flow" parameter of SLDF recordings as a dependent variable and all estimated CDI parameters as independent variables. Multiple linear regression was also performed, including the "flow" parameter of SLDF recordings as a dependent variable and PI of the CDI parameters of all the measured blood vessels as independent variables. RESULTS: The "velocity" parameter of SLDF was significantly correlated with the PI in the central retinal artery (P=0.02), PI and RI in the central retinal vein (P=0.01; P=0.01) and the PSV, MV, PI and RI in the short posterior ciliary artery, as recorded by CDI (P=0.003; P=0.02; P=0.002; P=0.01). The "volume" parameter of SLDF was significantly correlated with the PI and RI in the central retinal vein (P=0.03; P=0.03) and the PSV in the short posterior ciliary artery (P=0.03), as recorded by CDI. The "flow" parameter of SLDF was significantly correlated with the PI and RI in the central retinal vein (P=0.01; P=0.01) and the PSV, MV, PI and RI in the short posterior ciliary artery (P=0.003; P=0.03; P=0.002; P=0.01) as measured by CDI. The multiple regression analysis was statistically non-significant (P=0.86). The multiple linear regression analysis indicated that from among the PI of the evaluated blood vessels, the PI of the short posterior ciliary artery was the most significant predictor of the "flow" parameter of SLDF (P=0.01). CONCLUSION: This study suggests a positive correlation between the recordings of SLDF of the retinal tissue and the retrobulbar circulatory parameters of the CDI in diabetic patients without diabetic retinopathy.

Blood Flow Velocity↗
PubMed1996

A rapid dual organ rat carcinogenesis bioassay for evaluating the chemoprevention of breast and colon cancer.

In this study we evaluated the effect of dietary administration of a high fat, low fiber diet (HRD) with or without 2% phytic acid (PA) on the development of mammary cancer and/or colon cancer in rats exposed to methylnitrosourea (MNU), azoxymethane (AOM) or MNU + AOM. The rats were fed a HRD alone or a HRD + 2% PA. At the end of week 2, the rats were given either a s.c. injection of MNU (50 mg/kg body wt) or one of normal saline (vehicle). At the end of weeks 3 and 4, the rats were given either a s.c. injection of AOM (15 mg/kg body wt per week) or one of normal saline (vehicle). Nine weeks after the injection of MNU or saline, 10 rats from each group were sacrificed and the mammary tumor incidence and the number of colonic aberrant crypt foci (ACF) were compared between different groups. The administration of different diets was continued for an additional 21 weeks and the mammary tumor and colon tumor incidence between different groups were compared. Results showed that rats injected with MNU alone did not develop ACF or colon tumors while those injected with AOM alone did not develop mammary tumors. Linear regression analysis of the number of ACF at 11 weeks versus colonic tumor incidence at 32 weeks, and the linear regression analysis of mammary tumor incidence at 11 weeks versus mammary tumor incidence at 32 weeks, both showed good linear correlation. These results demonstrate the potential value of the short term dual organ carcinogenesis bioassay for screening chemopreventive agents for their relative ability to inhibit the development of mammary cancer and/or colon cancer while on high risk diet.

Animals↗
PubMed2005

On the effect of misclassification on bias of perfectly measured covariates in regression.

This note clarifies under what conditions a naive analysis using a misclassified predictor will induce bias for the regression coefficients of other perfectly measured predictors in the model. An apparent discrepancy between some previous results and a result for measurement error of a continuous variable in linear regression is resolved. We show that similar to the linear setting, misclassification (even when not related to the other predictors) induces bias in the coefficients of the perfectly measured predictors, unless the misclassified variable and the perfectly measured predictors are independent. Conditional and asymptotic biases are discussed in the case of linear regression, and explored numerically for an example relating birth weight to the weight and smoking status of the mother.

Bias↗
PubMed2006

Time trends of incidence of digestive system cancers in changle of China during 1988-2002.

AIM: To analyze the incidence of digestive system cancer in Changle of China over a 15-year period. METHODS: The datasets were presented as time-series of China-standardized annual incidence during 1988-2002. Linear regression model was used to analyze the incidence of stomach, liver, esophagus and colorectal cancers. RESULTS: Linear regression models for the time-series of stomach and esophagus cancer incidences for both men and women were statistically significant (P < 0.05); Regression models for liver cancer and for colorectal cancer were statistically significant for men (P < 0.05). CONCLUSION: The incidence rates of stomach and esophagus cancers for both men and women had down tendencies. For men, liver cancer had a down trend of the incidence and colorectal cancer had an upward trend of the incidence rate.

China↗
PubMed2007

Reducing vascular variability of fMRI data across aging populations using a breathholding task.

The magnitude and shape of blood oxygen level-dependent (BOLD) responses in functional MRI (fMRI) studies vary across brain regions, subjects, and populations. This variability may be secondary to neural activity or vasculature differences, thus complicating interpretations of BOLD signal changes in fMRI experiments. We compare the BOLD responses to neural activity and a vascular challenge and test a method to dissociate these influences in 26 younger subjects (ages 18-36) and 24 older subjects (ages 51-78). Each subject performed a visuomotor saccade task (a vascular response to neural activity) and a breathholding task (vascular dilation induced by hypercapnia) during separate runs in the same scanning session. For the saccade task, signal magnitude showed a significant decrease with aging in FEF, SEF, and V1, and a delayed time-to-peak with aging in V1. The signal magnitudes from the saccade and hypercapnia tasks showed significant linear regressions within subjects and across individuals and populations. These two tasks had weaker, but sometimes significant linear regressions for time-to-peak and coherence phase measures. The significant magnitude decrease with aging in V1 remained after dividing the saccade task magnitude by the hypercapnia task magnitude, implying that the signal decrease is neural in origin. These findings may lead to a method to identify vascular reactivity-induced differences in the BOLD response across populations and the development of methods to account for the influence of these vasculature differences in a simple, noninvasive manner.

Adolescent↗
PubMed2004

Signal-averaged P wave duration and the dimensions of the atria.

BACKGROUND: Delay of atrial electrical conduction measured as prolonged signal-averaged P wave duration (SAPWD) could be due to atrial enlargement. Here, we aimed to compare different atrial size parameters obtained from echocardiography with the SAPWD measured with a signal-averaged electrocardiogram (SAECG). METHODS: In 74 patients scheduled for elective echocardiography, an SAECG was recorded directly after the echocardiogram. We measured the SAPWD and registered clinical characteristics. The correlation between the SAPWD and the left atrial diameter (LAD), left atrial volume (LAV), right atrial volume (RAV), and total atrial volume (TAV) was analyzed by linear regression analyses. The effect of concomitant risk factors on TAV and the SAPWD was examined. RESULTS: Linear regression analysis showed that the correlation between the SAPWD and the LAD was significant (R(2)= 0.11, P = 0.03). However, LAV (R(2)= 0.15, P = 0.009), RAV (R(2)= 0.27, P = 0.0003), and TAV (R(2)= 0.37, P < 0.0001) were more strongly correlated to the SAPWD. The TAV and the SAPWD were not significantly associated with coexisting risk factors. CONCLUSIONS: The SAPWD is significantly correlated to the atrial size; most strongly to the TAV. The size of the right atrium, with the sinus node area, appears to affect the SAPWD.

Adult↗
PubMed2002

Influencing factors on chronic endothelial cell loss characterised in a homogeneous group of patients.

BACKGROUND/AIM: Advanced donor age, long death to excision time interval, and factors related to organ culture can trigger unfavourable intracellular processes in the graft endothelium and contribute to chronic endothelial cell loss after penetrating keratoplasty. The aim of this study was to investigate factors influencing chronic endothelial cell loss in a homogeneous group of patients. METHODS: 177 patients after first normal risk keratoplasties for keratoconus were retrospectively selected from the quality control database of our clinic. For 71 of them at least four central endothelial cell density values were documented in follow up. From these patients, only those 53 without any further intraocular procedures, without glaucoma, and without graft rejection were considered. A scatter plot of logarithmised endothelial cell density values against postoperative time was drawn for each patient. The slope of the regression line then equals the constant of decay in central endothelial cell density. The influence of donor age and storage time in organ culture on this index value of cell loss was investigated by means of linear regression analysis. RESULTS: Mean loss of central endothelial cell density was 16.7% per year. Regression analysis revealed a statistically significant negative linear effect of both postmortem time (beta = -0.324; p = 0.014) and donor age (beta = -0.282; p = 0.036) and a trend for storage time in organ culture (beta = -0.195; p = 0.142) in a combined linear regression model. CONCLUSION: Increased postmortem time and advanced donor age exert a significant negative effect on chronic endothelial cell loss. Storage time in organ culture seems to be third influencing factor. These negative influences may be reduced by compensating advanced donor age with minimised postmortem and storage time.

Age Factors↗
PubMed2003

Single-beat determination of Doppler-derived aortic flow measurement in patients with atrial fibrillation.

The clinical assessment of left ventricular systolic function in patients with atrial fibrillation is unreliable and difficult because of beat-to-beat variation. We initially evaluated an index that is on the basis of the ratio of preceding R-R (RR1) to pre-preceding R-R (RR2) intervals (RR1/RR2) for the measurement of Doppler aortic flow (peak flow velocity [Vp] and time-velocity integral [TVI] proportional to stroke volume) in 20 patients (aged 65 +/- 9.6 years) with atrial fibrillation. We obtained each parameter for >13 cardiac cycles, and the relationship between each parameter at a given cardiac beat and the RR1/RR2 ratio were evaluated by linear regression analysis. The value of each parameter at RR1/RR2 = 1 was calculated from the equation of linear regression line and compared with measured average value over all cardiac cycles. Both parameters showed a significant positive correlation with the RR1/RR2 ratio (Vp, r = 0.98, y = 1.01x + 0.61; TVI, r = 0.99, y = 1.01x + 0.26). The calculated value of each parameter at RR1/RR2 = 1 was quite similar to the average value (Vp, 97.4 +/- 30.8 vs 95.7 +/- 29.8 cm/s; TVI, 17.7 +/- 6.8 vs 17.3 +/- 6.7 cm, respectively). In the additional 20 patients (aged 77.4 +/- 15.2 years), Doppler aortic flow parameters of a single beat with identical RR1 and RR2 intervals were compared with measured average value over all cardiac cycles and showed similar results (Vp, r = 0.99, y = 0.99x + 3.4, P <.0001, bias -0.5 cm/s; TVI, r = 0.99, y = 0.92x + 1.5, P <.0001, bias 0.1 cm). In conclusion, the Doppler aortic flow at RR1/RR2 = 1 allows the left ventricular systolic parameters to be accurately evaluated during atrial fibrillation and obviates the less reliable process of averaging multiple irregular beats.

Aged↗
PubMed1992

Influx in vivo of low density, intermediate density, and very low density lipoproteins into aortic intimas of genetically hyperlipidemic rabbits. Roles of plasma concentrations, extent of aortic lesion, and lipoprotein particle size as determinants.

To compare the atherogenic potential of low density lipoprotein (LDL), intermediate density lipoprotein (IDL), and very low density lipoprotein (VLDL) under conditions where plasma levels of these lipoproteins are elevated, the influx of cholesterol in these lipoproteins into the aortic intima was measured in vivo in genetically hyperlipidemic rabbits from the St. Thomas's Hospital strain, an animal model that shares many of the features of the human disorder familial combined hyperlipidemia. Univariate linear regression showed that the arterial influx of LDL cholesterol (n = 25), IDL cholesterol (n = 14), and VLDL cholesterol (n = 10) was positively and linearly associated with plasma concentrations of LDL cholesterol in the range 0.2-6.4 mmol/l, of IDL cholesterol in the range 0.1-7.0 mmol/l, and of VLDL cholesterol in the range 0.7-8.5 mmol/l, respectively, and also with the extent of lesions in the arterial intima in the range 0-100% of the surface area. Multiple linear regression suggested that the arterial influx of LDL, IDL, and VLDL cholesterol was linearly dependent on plasma concentration, independent of lesion size. Furthermore, it appeared that the arterial influx of the three lipoproteins was linearly dependent on the extent of the lesions, independent of lipoprotein concentration. When influx was normalized for plasma concentration (intimal clearance) and for lesion size (compared within the same aorta), the intimal clearance of the larger IDL and VLDL particles was 15-35% less than that of the smaller LDL particles. These findings suggest that the quantitatively most important mechanism for transfer of plasma lipoproteins into the arterial intima involves nonspecific molecular sieving and that at elevated plasma levels, IDL and VLDL share with LDL the potential for causing atherosclerosis.

Animals↗
PubMed1999

Assessment of various parameters in the estimation of differential renal function using technetium-99m mercaptoacetyltriglycine.

Differential renal function (DRF) is an important parameter that should be assessed from virtually every dynamic renogram. With the introduction of technetium-99m mercaptoacetyltriglycine (99mTc-MAG3), a tracer with a high renal extraction, the estimation of DRF might hopefully become accurate and reproducible both between observers in the same institution and also between institutions. The aim of this study was to assess the effect of different parameters on the estimation of DRF. To this end we investigated two groups of children: group A, comprising 35 children with a single kidney (27 of whom had poor renal function), and group B, comprising 20 children with two kidneys and normal global function who also had an associated 99mTc-dimercaptosuccinic acid scan (99mTc-DMSA). The variables assessed for their effect on the estimation of DRF were: different operators, the choice of renal regions of interest (ROIs), the applied background subtraction, and six different techniques for analysis of the renogram. The six techniques were based on: linear regression of the slopes in the Rutland-Patlak plot, matrix deconvolution, differential method, integral method, linear regression of the slope of the renograms, and the area under the curve of the renogram. The estimation of DRF was less dependent upon both observer and method in patients with two normally functioning kidneys than in patients with a single kidney. The inter-observer comparison among children in either group was not dependent on either ROI or background subtraction. However, in patients with poor renal function the method of choice for the estimation of DRF was dependent on background subtraction, though not ROI. In children with two kidneys and normal renal function, the estimation of DRF from the 24 techniques gave similar results. Methods that produced DRF values closest to expected results, from either group of children, were the Rutland-Patlak plot and matrix deconvolution methods.

Area Under Curve↗
PubMed1987

The modulation of long-term potentiation by delta-9-tetrahydrocannabinol in the rat hippocampus, in vitro.

The duration of long-term potentiation (LTP) of the CA1 evoked field potential in rat hippocampal slices was significantly modulated by pre-treatment of slices with delta-9-tetrahydrocannabinol (THC) added to the incubation media. The three THC doses tested: 10 picomolar (pM), 100 pM, and 1000 pM, resulted in a biphasic change in population spike amplitude, such that 10 pM resulted in an increase, while 100 and 1000 pM resulted in dose-related decreases as compared to the control treatment. Upon subsequent induction of LTP by tetanizing stimulation, the THC treatments resulted in significant changes in the duration but not magnitude of potentiation. The early component of potentiation, post-tetanic potentiation, or PTP, was not affected by the THC treatments. LTP was seen to decay in an exponential manner over the 121 mini post-tetanus monitoring period. For comparisons of LTP duration, therefore, the half-life (t1/2) of LTP was extrapolated from linear regression analysis. The t1/2 values were determined for each treatment group from the slopes of the linear regression analysis of the logarithmically transformed time course data. While the control group t1/2 was determined as 280 min, the t1/2 for the THC groups were: 350 min (10 pM); 91 min (100 pM); and 33 min (1000 pM) doses, respectively. In context with previous reports of the disruptive action of marihuana intoxication on learning and memory, these results suggest that one possible action of THC may be in modulating hippocampal electrophysiology and its role in short-term memory processes.

Animals↗
PubMed1998

Determination of hexazinone and its metabolites in groundwater by capillary electrophoresis.

A micellar electrokinetic chromatographic method was developed to separate and quantify hexazinone and metabolites C, A1, E, B and D in groundwater. Hexazinone and its metabolites were extracted from water using Supelclean ENVI-Carb solid-phase extraction tubes. Quantitation was performed using UV photodiode detection at 220, 225, 230 and 247 nm. Intra-assay and inter-assay reproducibility studies run at 0.5, 1.0, 2.0 and 5.0 ppb indicated the procedure was reproducible. Groundwater samples collected from US Geological Survey monitoring wells were analyzed for hexazinone and its metabolites by CE. A comparison was made between CE and an established HPLC method of the hexazinone and metabolite B. The linear regression for hexazinone was y = 1.007x + 0.219 with a correlation coefficient of 0.96 while the linear regression for metabolite B was y = 1.100x-0.057 with a correlation coefficient of 0.91.

Electrophoresis, Capillary↗
PubMed2006

A semiquantitative method for the determination of reactive metabolite conjugate levels in vitro utilizing liquid chromatography-tandem mass spectrometry and novel quaternary ammonium glutathione analogues.

An in vitro semiquantitative reactive metabolite detection assay is described that incorporates NADPH-supplemented human liver microsomes, a novel quaternary ammonium glutathione analogue conjugating agent (QA-GSH), and liquid chromatography-tandem mass spectrometry (LC-MS/MS) for detection. The assay was developed to have high sample capacity and the potential for high sample throughput. MS/MS detection is selective and sensitive for the QA-GSH conjugating agent and semiquantitation of QA-GSH-reactive metabolite conjugates is performed using QA-GSH standards added to samples prior to analysis [i.e., internal standards (ISs)]. The reactive metabolite trapping capability of the free thiol group in QA-GSH was assessed using model drugs acetaminophen, clozapine, and flutamide, which are bioactivated to afford reactive metabolites. MS signal responses of equimolar amounts of QA-GSH standards were compared to assess the feasibility of using a QA-GSH IS approach to semiquantify reactive metabolite levels in vitro. The full scan Q1 MS response for each standard was within 3.3-fold of one another even though the "parent" moiety structure of each QA-GSH conjugate standard differed significantly. Standard curve analysis using selected reaction monitoring for each QA-GSH standard gave slope values that differed by only 1.5-fold. The QA-GSH IS semiquantitation method was tested by determining the level of QA-GS-acetaminophen conjugate formation at three different concentrations of acetaminophen and comparing the results to those from linear regression of authentic standards. The calculated levels of conjugate formed compared closely with those calculated from linear regression data of authentic standard curves. These results show that the QA-GSH semiquantitation assay described herein is a viable method for semiquantitatively assessing the bioactivation potential in vitro and is well-suited for use in early drug discovery high throughput screening paradigms.

Acetaminophen↗
PubMed1996

Setting targets for CABG surgery in the north western region.

BACKGROUND: A national target for coronary artery bypass graft (CABG) operations of 300 per million population was established by the Department of Health in 1987. Regional Health Authorities were required in 1993-1994 to achieve this target. The 1990-1991 Regional all-ages CABG rate was 239 per million population. The North Western Region planned to devolve the purchasing of certain services, including cardiac surgery, to districts in April 1993. Three districts had provider units within their boundaries. No means existed for estimating the appropriate operation rate, according to need, for district populations. A method was therefore devised for calculating guideline targets. METHODS: Based on the assumption that the ischaemic heart disease standardized mortality ratio (IHD SMR) is a valid proxy of the need for CABG operations and using a linear regression model, the existing district IHD SMRs and IHD hospitalization rates were used to derive guideline targets for districts, taking account of the need to raise the average operation rate for the Region to 300 per million population. Further targets using alternative Poisson regression model and weighting formulae were also derived. RESULTS: The derived operation targets based on the linear regression model were lower than the existing operation rate in the three provider districts and in the biggest provider's two neighbours. In all other districts, except one, targets were higher than the existing number of operations. The two further analyses produced essentially similar results. CONCLUSIONS: In the North Western Region wide variations in CABG operation rates existed. Access to CABG surgery in the North Western Region appears to depend on proximity to the provider districts. Operation rates should reflect need. This paper presents a method for deriving targets which take account of need.

Adult↗
PubMed2006

Waiting times for cataract surgery in ten European countries: an analysis using data from the SHARE survey.

AIMS: To assess waiting times for cataract surgery and their acceptance in European countries, and to find explanatory, country-specific health indicators. METHODS: Using data from the survey of health, ageing and retirement in Europe (SHARE), waiting times for cataract surgery of 245 respondents in ten countries were analysed with the help of linear regression. The influence of four country specific health indicators on waiting times was studied by multiple linear regression. The influence of waiting time and country on the wish to have surgery performed earlier was determined through logistic regression. Additional information was obtained for each country from opinion leaders in the field of cataract surgery. RESULTS: Waiting times differed significantly (p<0.001) between the ten analysed European countries. The length of wait was significantly influenced by the total expenditure on health (p<0.01) but not by the other country specific health indicators. The wish to have surgery performed earlier was determined by the length of wait (p<0.001) but not by the country where surgery was performed. CONCLUSION: The length of wait is influenced by the total expenditure on health, but not by the rate of public expenditure on health, by the physician density or by the acute bed density. The wish to have surgery performed earlier depends on the length of wait for surgery and is not influenced by the country.

Aged↗
PubMed1991

Oxy- and carboxyhemoglobin saturation determination in frozen small vessels.

A four-wavelength microspectrophotometric method has been developed to determine both O2 and CO saturation of hemoglobin in frozen small vessels. This method compensates for the nonspecific light loss related to the ice crystals and in addition can simultaneously measure both hemoglobin O2 and CO saturation in frozen blood. By using four wavelengths in the 540-580 nm range, it is possible to determine both O2 and CO saturations in frozen blood. This four-wavelength method demonstrated an accuracy of 6-8% within 95% confidence limits (SE of estimate = 3-4%) for both CO and O2 saturation determinations in frozen blood. The linear regression of the hemoglobin O2 saturation calibration was significant with an r2 value of 0.989. The linear regression of the CO saturation calibration was also significant with an r2 of 0.996. The repeatability of each measurement was on the order of 1-6%. This method also represents an accuracy within 5% for measurements in the frozen rabbit heart for both CO and O2 in vessels between 20 and 100 microns in diameter. It compares favorably with other previously reported methods. It shows a similar accuracy but requires fewer repeat measurements than the three-wavelength method. This method can provide accurate measurement of O2 and CO saturation in small frozen arterioles and venules of any organ.

Animals↗
PubMed2004

Computerized cardiotocography parameters in pregnant women affected by pregestational diabetes mellitus.

AIM: To evaluate whether computerized CTG (cCTG) is a reliable method of predicting neonatal outcome in pregnancies complicated by pregestational diabetes at term. PATIENTS AND METHODS: We considered 27 pregnant women affected by pregestational diabetes and 46 normal pregnancies as controls that fulfilled the following criteria: singleton, Caucasian, euglycemic pregnancies at term (>37 weeks gestational age). All women delivered by cesarean section (CS), with an antepartum cCTG performed within one hour before the CS and an UBGA available at birth. No patient was in labor during FHR monitoring. RESULTS: Among cCTG parameters, accelerations 15 bpm, HV min, HV ms and STV were significantly lower in comparison to controls. We observed that in the diabetic pregnant women the parameter STV was not able to predict or to linearly regress with the most important UBGA parameters: pH and pCO2. Contrarily, in normal pregnancies, the STV linearly regressed with both the pH (p < 0.03) and pCO2 (p<0.04). CONCLUSIONS: Computerized FHR criteria may not be applicable to fetuses in pregestational diabetic pregnancies at term. Therefore some criteria should perhaps be modified for a correct interpretation of cCTG in these pregnancies.

Acidosis↗
PubMed2003

Sonography of polyethylene liners used in total knee arthroplasty.

OBJECTIVE: We investigated the ability of sonography to reveal the polyethylene liner used in total knee arthroplasty with the hopes of establishing a possible relationship between the sonographic measurement of the actual thickness of the polyethylene liner and the estimated thickness based on conventional radiography. MATERIALS AND METHODS: Twenty-four consecutive patients who were referred for Doppler screening for deep venous thrombosis after total knee arthroplasty were evaluated. The polyethylene liner was identified on sonography, and three measurements were obtained from four locations: anteromedial joint line (just medial to the midline incision), along the medial joint line, anterolateral joint line (just lateral to the midline incision), and along the lateral joint line. These sonographic measurements were compared with radiographic measurements of the radiolucent polyethylene liner and with the manufacturers' stated size of the polyethylene liner. Linear regression analyses were then performed. RESULTS: The polyethylene liner is seen on sonography as a strong linear echogenic interface with posterior acoustic shadowing. Linear regression analyses showed a high correlation (r = 0.8) between the sonographic measurements and the radiographic measurements. A relatively poor correlation (r = 0.2) was noted between the manufacturers' stated size of the liner and the sonographic measurements. CONCLUSION: We found that the polyethylene liner used in total knee arthroplasty can be clearly identified during sonographic evaluation of the knee. We also found a high correlation between the longitudinal measurement of the polyethylene liner with the thickness of the radiolucent polyethylene as measured on conventional radiographs. We propose that sonography could potentially be a useful noninvasive imaging modality to screen for subtle cases of polyethylene wear.

Aged↗