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 451 records · Page 25Linked to original sources
PubMedDate not supplied

[Factors influencing variability of PSA blood concentration in patients with prostate benign hyperplasia].

Study conducted in 107 patients with prostate benign hyperplasia to establish the influence that prostate volume, age, sexual activity, extent of obstruction and presence of urethral catheter have on PSA serum levels, using a linear regression model. In simple linear regression studies, a good correlation was seen between PSA serum concentration and prostate volume calculated by transrectal ultrasound (r = 0.57), which decreases when age in used as an independent variable (r = 0.40). In multiple linear regression studies where all the variables are included, correlation increases slightly (r = 0.60), and only prostate volume and sexual activity are significant. Prostate volume in the single most influential factor on PSA levels in prostate benign hypertrophy although it is not the only parameter to be taken into account in serum level variations of this marker. Age influence can be secondary to other factors such as sexual activity.

Aged↗
PubMed2000

A semiparametric approach to analysing dose-response data.

In the analysis of a quantal dose-response experiment with grouped data, the most commonly used parametric procedure is logistic regression, commonly referred to as 'logit analysis'. The adequacy of the fit by the logistic regression curve is tested using the chi-square lack-of-fit test. If the lack-of-fit test is not significant, then the logistic model is assumed to be adequate and estimation of effective doses and confidence intervals on the effective doses can be made. When the tolerance distribution of the dose-response data is not known and cannot be assumed by the user, one can use non-parametric methods, such as kernel regression or local linear regression, to estimate the dose-response curve, effective doses and confidence intervals. This research proposes another alternative based on semi-parametric regression to analysing quantal dose-response data called model-robust quantal regression (MRQR). MRQR linearly combines the parametric and non-parametric predictions with the use of a mixing parameter. MRQR uses logistic regression as the parametric portion of the model and local linear regression as the non-parametric portion of the model. Our research has shown that the MRQR procedure can improve the fit of the dose-response curve by producing narrower confidence intervals for predictions while providing improved precision of estimates of the effective doses with respect to either logistic or local linear regression results.

Dose-Response Relationship, Drug↗
PubMed2001

Relationships among scrotal and testicular characteristics, sperm production, and seminal quality in 129 beef bulls.

Standard breeding soundness examinations plus measurement of scrotal surface temperature (SST), internal/scrotal testicular temperatures, testicular ultrasonographic echotexture, daily sperm production, and epididymal sperm reserves were conducted on 129, 16-month-old crossbred beef bulls. There were significant positive linear correlations between SST and internal scrotal/testicular temperatures, a positive linear regression (P < 0.06) of bottom SST with the incidence of secondary sperm defects, but a negative linear regression (P < 0.01) with the incidence of primary sperm defects. Testicular echotexture had a positive linear regression with daily sperm production (P < 0.002) and testicular tone had a negative linear regression (P < 0.008) with epididymal sperm reserves. Scrotal circumference had a positive linear regression (P < 0.04) with the percentage of progressively motile sperm, a negative linear regression (P < 0.1) with the incidence of primary sperm defects, and a positive linear regression (P < 0.0001) with epididymal sperm reserves. In addition to seminal quality and scrotal circumference, testicular ultrasonographic echotexture has considerable promise for augmenting breeding soundness examinations of bulls.

Animals↗
PubMed1993

[Usefulness of residuals in clinical research].

The simple linear regression analysis, multiple linear regression and logistic regression constitute powerful statistical analysis tools widely used in clinical research. These kinds of analyses are based upon mathematical models which at the same time are established on certain basic assumptions. The regression analysis assumptions are basically: a) that the model is really linear, b) that the distribution of data is normal (from a statistical point of view), c) that the variances of the employed data are homogeneous (homocedastics) and that the included data are independent. The regression diagnostic has become popular as a form to evaluate if the assumptions have been accomplished, one of its most important techniques is the residual analysis. A residual can be defined as the value which measures the distance between the regression line and the corresponding value of the variable "y". Among these kinds of residuals used to evaluate the assumptions of regression are: the crude residual, the standardized, of student and the jackknife. The most useful among them is the jackknife residual. The usefulness and limitations of the residuals in the evaluation of the regression analysis assumptions are described, basically referring to the identification and handling of extreme values (outliers).

Evaluation Studies as Topic↗
PubMed2002

A genotypic drug resistance interpretation algorithm that significantly predicts therapy response in HIV-1-infected patients.

OBJECTIVES: The development of a genotypic drug resistance interpretation algorithm, and the evaluation of its power to predict therapy outcome. DESIGN: A rule-based algorithm was established by an individual expert and was based on published and in-house results, independently from the data of the patients used in this evaluation. The predictive value of the algorithm for virological outcomes was retrospectively evaluated using the baseline genotype observed in patients on highly active antiretroviral therapy, failing virologically and subsequently starting a salvage regimen. METHODS: The independent association between the susceptibility score (calculated according to the algorithm) and the virological response at 3 months, was analysed using multivariable logistic regression and multiple linear regression models. RESULTS: In two clinical centres 240 patients were studied. At 3 months 35% had a viral load of <500 RNA copies/ml. Using multivariable logistic regression, the odds ratio of achieving a viral load <500 RNA copies/ml at month 3 per unit increase of susceptibility score was 2.0 (95% CI 1.3-3.1; P=0.002) after adjusting for baseline viral load, genotype-driven salvage therapy, number of new drugs in the regimen, use of a new drug class in the regimen, nelfinavir-containing salvage therapy and history of prior viral load <500 RNA copies/ml. Using multiple linear regression, the susceptibility score showed a significant linear correlation with the log viral load change (slope=-0.27 log10 RNA copies/ml; 95% CI -0.11 to -0.43; P=0.001) after adjusting for history of prior viral load <500 RNA copies/ml, number of new drugs in the salvage therapy, use of a new drug class in the salvage therapy and baseline viral load. CONCLUSIONS: This algorithm proved to be a significant independent predictor of therapy response at 3 months in this cohort of HIV-1-infected patients on salvage therapy. However, it should be subject to regular updates as is needed in this fast developing field.

Adolescent↗
PubMed2005

[Estimation of CO2 fluxes from rice paddies based on transparent chamber measurement].

Closed chamber provide a valuable tool for measuring CO2 exchange fluxes. In general, the change rate of CO2 concentration is assumed to be constant in the short measurement time and a linear regression method is used to estimate the CO2 fluxes. However, due to the physical and physiological effects, the change rate of CO2 is not always constant. A linear regression method and an exponential regression method are compared in this study. Results show that during the growing stages except for the ripening, nonlinear relationship of CO2 concentration versus time was estimated in the sunny daytime. Absolute values of CO2 fluxes calculated by linear regression were lower than that by exponential regression. Whereas, CO2 concentration changed linearly with time in cloudy days or in nighttime. And no significant difference was found between the results calculated with these two methods. Accumulative CO2-C fluxes with exponential and linear regressions were compared with the values of net ecosystem exchange of CO2-C (NEE), which were calculated by net primary production (NPP) minus soil respiration. The values with the exponential regression method were closer to NEE than those with linear regression. Therefore, the linear regression method could result in underestimation of carbon budget of ecosystem. Based on transparent chamber measurement, the exponential regression is more feasible to calculate CO2 fluxes than the linear regression.

Carbon Dioxide↗
PubMed1984

Determination of the physical working capacity in children using three different regression models.

Experimental findings of the working capacity at a heart rate of 170 bts/min (W170) were compared to predicted values. Statistical tests were applied to examine the suitability and the error of prediction of three different regression models: a linear regression line, a polynomial regression model, and a "break point" regression model, which were compared to the time course of the heart rate during a linearly increasing work load from 0 to 100 W during 10 min. For this study the results of 28 children, 15 and 16 years old, and students of physical education were investigated. When a linear regression line was compared to these data, systematic deviations between measured data and the values estimated by this model were found. When the W170 was predicted using this model from the data collected during the first 10 min of an exercise procedure for the determination of the heart rate index, the physical working capacity was overestimated. The polynomial regression model and the "break point" regression model agreed with the time course of the heart rate without systematic error and allowed an unbiased prediction of the W170 from the first 10 min of the exercise test.

Adolescent↗
PubMed2005

[Study on the current situation and influential factors of child neglect among aged 3-6 year-olds in the urban areas of China].

OBJECTIVE: Children aged 3 - 6 years in the urban areas of China were surveyed for the first time to find out the state of child neglect (CN) as well as the major relevant risk factors so as to provide evidence for developing intervention measures. METHODS: 1163 children (of whom 49.6% were males and 4.5% were minority nationality) were randomly sampled under multistage stratification, from 25 cities which representing 15 provinces of China. Based on the Child Neglect Norms used by China, prevalence of CN was identified and SPSS-Windows 11.0 was employed for statistical analysis. Scores, frequency/degrees, age, sex and 5 types (physical, emotional, educational, medical and safety) of CN on every group of the regions, were calculated. Multifactorial analysis was conducted through Binary Logistic Regression and multiple linear regression to determine the relevant risk factors. RESULTS: (1) The average degree of CN for the 3 - 6 year-olds was 42.2, with its prevalence as 28.0%. Degrees of CN for the groups of 3, 4, 5, 6-year-olds were 41.7, 42.2, 42.1 and 43.1 (F = 0.988, P > 0.05), with frequencies of 25.0%, 25.3%, 27.9% and 35.4% (chi(2) = 4.798, P > 0.05), respectively. Degrees for CN in males and females were 42.7 and 41.8 (F = 2.502, P > 0.05) with the frequencies as 32.6% and 23.7% (chi(2) = 6.585, P < 0.05), respectively. Degrees of CN for the five types were 39.4-43.4 with the frequencies as 5.1%-12.9%, respectively. No significant difference was found in the frequency of the types (with an exception on 'physical neglect') between males and females (P > 0.05). The highest frequency (42.9%) of CN was seen in the single-parent families and the lowest in large family with three generations (25.5%). (2) According to monofactorial chi(2) test, the possible risk factors of CN would include: educational background, occupation and decrease of income of the parents during last year, etc. (3) Binary Logistic regression analysis showed that the influential factors to the occurrence of CN would include: father's educational background, sex of the child and mother's occupation, etc. (4) Multiple linear regression showed that the influential factors to the degree of CN were: family structure, number of supporting family members, relationship between parents and children, etc. CONCLUSION: The degree and frequency of CN among children aged 3 to 6 in the urban areas of China were high but similar among the four age groups. Male children had a higher frequency of neglect than females, but with similar degree. Children in single-parent families had the highest frequency. The major influential factors of CN would include: educational background, occupation, family structure, family income of the parents which were similar to the results reported from foreign literature.

Child↗
PubMed2005

The role of climate on Malaria Incidence Rate in four governorates of Yemen.

The aim of study was to investigate the role of climate on the Malaria Incidence Rates (MIR) in some regions in of Yemen. For such purpose, the monthly (MIR) were calculated from the records of the hospitals' laboratories and centers of the Malaria Rollback centers in the main cities of the governorates Hudeidah, Taiz, Sana'a and Hadramout for the period 1989-1998. The readings of the climatic factors (CF) particularly the average monthly temperature (T), relative humidity (RH), volume of rain fall (RF) and wind speed (WS) for the same period of time were also collected from different weather and climatic information resources. Descriptive statistics, simple linear regression and multiple linear regression techniques were used to analyse the relationship between MIR and CF. The analysis shows highly significant relationship between MIR and the CF in these regions of Yemen (p-value 0.001).

Climate↗
PubMedDate not supplied

Correlation of bottom-to-top surface microhardness and conversion ratios for a variety of resin composite compositions.

Knoop microhardness (KHN) and degree of conversion (DC) values were obtained from the bottom and top surfaces of 1-, 2- and 3-mm thick samples of three types of resin composite: an anterior microfill, an anterior hybrid and a posterior hybrid, all having differing filler size and loading but similar shade (A2) and basic monomeric content. Sample infrared spectra were obtained using attenuated total reflectance (ATR) in a Fourier transform infrared (FTIR) spectrometer. The samples were exposed using a 40-second exposure to a quartz-tungsten-halogen light source with an irradiance of approximately 560 mW/cm2. They were stored for 24 hours in complete darkness at 37 degrees C and 100% humidity prior to obtaining cured spectra and KHN readings. KHN andDC values were obtained from the same sample specimen made at similar surface depths, but separate groups were made for obtaining top and bottom values. Cure and hardness data were analyzed with one- and two-way ANOVAs followed by the Tukey-Kramer post-hoc test. Linear regression was also applied. Statistical testing was per. formed at a pre-set 0.05 level of significance. KHN and DC were significantly different according to composite type and depth (p=0.0001), with an interaction effect (p=0.0022). KHN, DC and corresponding bottom/top surface (B/T) ratios decreased with depth. Regression revealed a linear relationship between DC and KHN for each composite type, with no r2 less than 0.96. When B/T ratios were correlated, a B/T-KHN ratio of 0.80 corresponded to a narrow range of B/T-DC ratios (between 88 to 91%) for the three composites tested. When combining results from composite types, linear regression of B/T-DC and B/T-KHN produced a very predictable relationship (r2=0.959), for which a B/T-KHN ratio of 0.80 corresponded to a B/T-DC ratio of 0.90. As a measure of completeness of conversion, B/T-KHN was approximately 2.5 times more sensitive than the B/T-DC ratio. In summary, while KHN cannot be used to directly compare conversion of the different composites tested, the use of B/T ratios for both hardness and conversion resulted in a linear relationship independent of filler size or loading.

Barium↗
PubMedDate not supplied

Preoperative plasma fibrinogen levels in gastric cancer patients correlate with extent of tumor.

BACKGROUND/AIMS: The aim of the present study was to investigate the relationship between the preoperative plasma fibrinogen level and the extent of tumor involvement in gastric cancer patients. METHODOLOGY: Preoperative plasma fibrinogen levels of 354 patients who underwent gastric cancer surgery were quantified using an immunoassay. The relationships between the plasma fibrinogen level and other prognostic variables (tumor size, macroscopic and histological type, depth of tumor invasion, presence of lymph node involvement and distant metastasis) were then examined using univariate and multivariate linear regression analyses. RESULTS: The plasma fibrinogen level was significantly lower in patients with early gastric cancer than in those with advanced gastric cancer (312+/-6.7 vs. 361.9+/-97.0 mg/mL, p<0.001). A significant relationship existed between the preoperative fibrinogen levels and the presence of metastatic lymph nodes (320+/-78.6 vs. 352.6+/-94.1 mg/mL, p=0.001) and distant metastasis (338.2+/-89.5 vs. 396.9+/-128.3 mg/mL, p=0.013). Size of the tumors and depth of tumor invasion could predict elevated fibrinogen levels positively in both the univariate regression and multivariate linear regression analyses. CONCLUSIONS: These data suggest that the plasma fibrinogen level is a clinically important and useful marker of the extent of tumor progression in gastric cancer.

Adult↗
PubMed2002

[Comparison of three regression methods in the calculation of LA(50)].

OBJECTIVE: To search for an optimal regression method for LA(50) in burn management. METHODS: Three kinds of regression methods, i.e. linear regression, probability type probit regression and logic type probit regression. were employed to calculate the LA(50) and its 95% confidence interval (CI) from different burn data. The results produced by these three methods were compared with one another. RESULTS: Three methods were used to calculate LA(50) of one sample with the same, different or evidently different LA(50), but the order of LA(50) from high to low values appeared to be the same (P > 0.05). When the 95% CI of LA(50) was calculated by three methods, significantly different results were obtained (P < 0.01), especially the 95% CI by linear regression was evidently different to that produced by two other kinds of probit regression (P < 0.01), while there was no difference in the 95% CI between the two probit regression methods (P > 0.05). CONCLUSION: Different methods for the calculation of LA(50) and 95% CI came to different results. The LA(50) from different samples could only be compared as valid when the same calculation method was employed. Probit regression (logic type) seemed to be the best method in the calculation of LA(50) in terms of its simplicity, practicability and accuracy.

Adolescent↗
PubMed2001

Faculty triage shortens emergency department length of stay.

OBJECTIVE: To determine whether faculty triage (FT) activities can shorten emergency department (ED) length of stay (LOS). METHODS: This was a comparison study measuring the impact of faculty triage vs no faculty triage on ED LOS. It was set in an urban county teaching hospital. Subjects were patients presenting to the registration desk between 9 AM and 9 PM on 16 consecutive Mondays (August 2 to November 15, 1999). On eight Mondays, an additional faculty member was stationed at the triage desk. He or she was asked to expedite care by rapid evaluation orders for diagnostic studies and basic therapeutic interventions, and by moving serious patients to the patient care areas. He or she was not provided with detailed instructions or protocols. The ED LOS, time of registration (TIMEREG), inpatient admission status (ADMIT), x-ray utilization (XRAY), total patients registered each day between 9 AM and 9 PM (TOTREG), and patients who left without being seen (LWBS) were determined using an ED information system. The LOS was analyzed in relation to FT, ADMIT, and XRAY by the Mann-Whitney U test. The LOS was related to TIMEREG and TOTREG by simple linear regression. Stepwise multiple linear regression models to predict LOS were generated using all the variables. RESULTS: Patients without FT (n = 814) had a mean LOS of 445 minutes. Patients with FT (n = 920) had a mean LOS of 363 minutes. Mean difference in LOS was -82 minutes (95% CI = -111 to -53), a reduction of 18%. The LOS was also related to: ADMIT +203 minutes (95% CI = 168 to 238), TOTREG -2.7 min/additional patient registered (95% CI = -1.15 to -4.3), and TIMEREG +0.14 min/min since 9 AM (95% CI = 0.07 to 0.21). The LWBS was reduced by 46% with FT. In multiple regression analysis, ADMIT, FT, TIMEREG, and XRAY were all related to LOS, but the model explained only a small part of variance (adjusted R(2) = 0.093). The faculty cost is estimated to be $11.98/patient. CONCLUSIONS: Faculty triage offers a moderate increase in efficiency at this ED, albeit with relatively high cost.

Emergency Service, Hospital↗
PubMed2000

Limited sampling model for the estimation of pharmacokinetic parameters in children.

A limited sampling model (LSM) is proposed for the first-time assessment of pharmacokinetic parameters (area under the concentration-time curve (AUC), Cmax, and T1/2) in children after a single oral dose of drug. Three drugs were evaluated in this study. The LSM was developed for each drug from the data of 10 healthy adult volunteers. The relationship at selected time points between plasma concentration and the AUC or Cmax was evaluated by multiple linear regression. The multiple linear regression that gave the best correlation coefficient (r) for 3 sampling times versus AUC or Cmax was chosen as the LSM. Pharmacokinetic parameters generated using sparse sampling (3 blood samples) were compared with pharmacokinetic parameters generated using extensive sampling (>7 blood samples). The results indicated that a limited sampling model can be developed from adult data to estimate pharmacokinetic parameters in children with fair degree of accuracy.

Administration, Oral↗
PubMed2004

Intraindividual variation in serum retinol concentrations among participants in the third National Health and Nutrition Examination Survey, 1988-1994.

BACKGROUND: The biological variability in serum retinol concentrations has never been examined in a large sample, and its effect on population distribution estimates and the clinical assessment of vitamin A status is unknown. OBJECTIVE: We evaluated the biological CV of serum retinol and examined the effect of the CV on both population distribution estimates and clinical assessments of vitamin A status by using data from the third National Health and Nutrition Examination Survey, 1988-1994. DESIGN: We described the biological CV [(SD/x) x 100] and examined associations between the CV and other factors via multivariate analysis of variance and linear regression. We used linear regression to predict the mean retinol concentration from a single concentration and established 95% CIs for each participant. We estimated the adjusted prevalence of inadequate vitamin A status (retinol < 1.05 micromol/L) on the basis of the CIs. We estimated an uncertainty range for serum retinol concentrations for which the CIs included the established cutoff. RESULTS: The mean biological CV across all strata was 6.45%. The biological CV varied significantly between racial-ethnic groups (P < 0.05). Prevalence estimates of inadequate serum retinol concentrations were reduced after adjustment for the total variation, with an adjusted overall prevalence of 0.62% compared with an unadjusted prevalence of 2.63%. CONCLUSIONS: The actual population prevalence of inadequate vitamin A status may be 75% lower than the estimates previously reported. Confirmation of vitamin A status may be needed for persons in the United States with observed serum retinol concentrations near the recognized cutoff.

Adolescent↗
PubMed1995

Racial differences in nasal fossa dimensions determined by acoustic rhinometry.

Sixty acoustic rhinographs from subjects of three different ethnic groups (Caucasian [Europeans], Negro, and Oriental) were examined at baseline and after decongestion. The main parameters analysed were minimal cross-sectional area (MCA), the distance at which this occurred (D), nasal volume at 0-4 cm (Vol), mean cross-sectional area at 0-6 cm (MA), and the cross-sectional area at 10 points in the nose (0, 0.5, 1, 1.5, 2, 2.5, 3, 4, 5, and 6 cm) analysed as a series (A). Values from left and right were combined and mean values used. Analysis was carried out using multiple linear regression and grouped linear regression with analysis of covariance and, for A, multifactorial analysis of variance. For MCA, race was the main determining factor with Orientals and Caucasians significantly lower than Negroes: p<0.0001 (corrected means and 95% confidence intervals [c.i.]: Orientals: 0.63 cm2, 0.55-0.71 cm2; Caucasians: 0.69 cm2, 0.62-0.77 cm2; Negroes: 0.87 cm2, 0.79-0.95 cm2). Height alone correlated with D in the decongested state (p<0.0001); race as well as height in non-decongested noses (p = 0.018). There were significant racial differences in Vol in both decongested (p = 0.014), and non decongested noses (p<0.0001). In the non-decongested state MA was significantly different in all racial groups: p<0.0001 (corrected means and c.i.: Orientals: 3.89 cm2, 3.47-4.31 cm2; Caucasians: 4.67 cm2, 4.27-5.09 cm2; Negroes: 5.13 cm2, 4.72-5.53 cm2). In the decongested state there was a significant difference between Negroes and the other two groups (p = 0.015), and Orientals and Caucasians were a homogenous population. We conclude that race has a significant effect on acoustic rhinometry measurements and this needs to be taken into account.

Acoustics↗
PubMed2000

The measurement of plasma catecholamine and lactate thresholds: a comparison of methods.

Plasma lactate and catecholamine thresholds were calculated using three different variations of linear regression, an algorithmic linear regression method, a log-log transformation method and a semi-log method. A group of 18 male sports science students undertook an incremental test to exhaustion on a cycle ergometer. A 5-ml blood sample was drawn at rest, after 4 min of exercise and every 2 min thereafter until the cessation of the test. Lactate, adrenaline and noradrenaline concentrations were measured. Lactate threshold (Th1a), adrenaline threshold (ThA) and noradrenaline threshold (ThNA) were calculated using each of the three methods. The best fits of the methods were examined by comparing their standard error of estimates (SEEs). The algorithmic method demonstrated a higher SEE than the other two methods, but only for Th1a and ThNA. The power output for which each method calculated the thresholds demonstrated a main effect for method. Tukey post hoc tests showed that the algorithmic method produced significantly higher outputs than the other two methods, which did not differ significantly from one another. Comparison of these power outputs showed that Th1a and ThA differed significantly, regardless of method, there were no other significant differences. Plasma concentrations of lactate, adrenaline and noradrenaline showed that the values of Th1a and ThNA calculated by the algorithmic method were significantly higher than those calculated using the other two methods, which did not differ significantly from one another. The only significant difference for ThA was between the algorithmic and semi-log methods. Correlations between the power outputs at which each method calculated the thresholds varied greatly between methods, and were at best only moderate (r = 0.63). It was concluded that the algorithmic method was less powerful than the other two methods, and that Th1a, ThA and ThNA are not highly correlated.

Adult↗
PubMed1999

Collection and interpretation of plasma leptin concentration data in humans.

OBJECTIVE: To reassess the relationship between body fat and plasma leptin concentrations when a) replicate measures of leptin are made; b) energy intake is controlled at isoenergetic levels before the study; and c) body fat and percent body fat are measured with dual energy X-ray absorptiometry (DXA). RESEARCH METHODS AND PROCEDURES: Two separate studies were conducted. In the first study, four plasma samples were collected for measurement of leptin over 30 minutes on a single day in 43 lean and obese men and women. For the second study, plasma samples were collected on four consecutive days from a group of 50 lean and obese men and women. Percent body fat (and body fat mass) was related to log-transformed mean plasma leptin concentrations using linear regression analysis; multiple linear regression analysis was used to determine whether there was an effect of gender on this relationship, and the analysis of Choi was used to examine whether percent body fat or body fat mass better predicts plasma leptin concentrations. RESULTS: For the first study, percent body fat was highly correlated (r=0.96, p<0.0001) with log-transformed mean leptin concentrations. No difference in the relationship between leptin and percent body fat in men and women was detected. The second study confirmed this observation; the relationship between ln leptin and percent body fat was virtually identical (r=0.93, p<0.001). Analysis of the pooled data suggests that percent body fat is a better predictor of plasma leptin concentration than body fat mass. The use of multiple (as opposed to a single) leptin measurements did not significantly improve the leptin/body fat relationship. DISCUSSION: When robust body composition techniques and diet control measures are taken into consideration, the relationship between ln plasma leptin concentrations and percent body fat is not different in men and women.

Absorptiometry, Photon↗