Browsing by Author "Marinković, J. (7004611210)"

Purpose: Dietary factors have been suggested as having a major role in the etiology of prostate cancer. The purpose of this study was to investigate whether there are differences in the dietary factors between younger and older men with prostate cancer. Materials and methods: A case-control study was conducted in two towns of Serbia without provinces. The case group consisted of 101 patients with histologically confirmed prostate cancer and the control group of 202 hospital patients with no neoplastic disorders (for each case two controls were chosen). Cases and controls were individualy matched by age (± 2 years), time of hospital admission and place of residence. Univariate and multivariate logistic regression methods were used for result anlayses. Results: After adjustment for a variety of confounding variables, a significant relationship between prostate cancer and vitamin B 12 (odds ratio - OR = 4.77; 95% confidence interval - 95% C.I. = 1.74-13.07) and potassium (OR = 0.18; 95% C.I. = 0.07-0.49) was found in men ≤ 70 years old. In men > 70 years old daily intake of retinol and atocopherol were independently significantly related to prostate cancer (OR = 3.20; 95% C.I. = 1.38-7.43 and OR = 0.42; 95% C.I. = 0.18-0.99, respectively). Conclusion: Differences between younger and older men in the dietary factors related to prostate cancer (retinol and a-tocopherol) are in line with the existing literature data. The relationship of prostate cancer with vitamin B12 and potassium should be corroborated by other authors.

Purpose: Dietary factors have been suggested as having a major role in the etiology of prostate cancer. The purpose of this study was to investigate whether there are differences in the dietary factors between younger and older men with prostate cancer. Materials and methods: A case-control study was conducted in two towns of Serbia without provinces. The case group consisted of 101 patients with histologically confirmed prostate cancer and the control group of 202 hospital patients with no neoplastic disorders (for each case two controls were chosen). Cases and controls were individualy matched by age (± 2 years), time of hospital admission and place of residence. Univariate and multivariate logistic regression methods were used for result anlayses. Results: After adjustment for a variety of confounding variables, a significant relationship between prostate cancer and vitamin B 12 (odds ratio - OR = 4.77; 95% confidence interval - 95% C.I. = 1.74-13.07) and potassium (OR = 0.18; 95% C.I. = 0.07-0.49) was found in men ≤ 70 years old. In men > 70 years old daily intake of retinol and atocopherol were independently significantly related to prostate cancer (OR = 3.20; 95% C.I. = 1.38-7.43 and OR = 0.42; 95% C.I. = 0.18-0.99, respectively). Conclusion: Differences between younger and older men in the dietary factors related to prostate cancer (retinol and a-tocopherol) are in line with the existing literature data. The relationship of prostate cancer with vitamin B12 and potassium should be corroborated by other authors.

The objective of the work was to study blood pressure and some anthropometric characteristics in children. The study design included the prevalence study at Belgrade, Serbia and Montenegro. The study comprised 1651 subjects (809 boys and 842 girls) aged 7-14 years, that is, 2.6% of all relevant population. The average levels of systolic and diastolic blood pressures (SBP and DBF) were 113.4/70.3 mmHg in boys and 114.6/71.1 mmHg in girls. High SBP was present in 4.7% of boys and in 5.3% of girls. High DBP was found in 5.6% of boys and in 4.8% of girls. According to multivariate regression analysis, in boys SBP and DBP were significantly associated with age, body mass index (BMI) and subscapular skinfold, SBP was also associated with body weight, and DBP with triceps skinfold. In girls, SBP was significantly related to BMI, suprailiac skinfold and body height, and DBP was significantly associated with BMI, suprailiac and subscapular skinfolds. The results of the present study support the opinion that BMI is a significant predictor of blood pressure in children and point out to suprailiac skinfold in girls as a possible predictor of blood pressure.

One hundred and one patients with histologically confirmed prostate cancer and 202 hospital controls individually matched by age (± 2 years), hospital admittance and place of residence, were interviewed during the period 1990-94 in two towns in central Serbia (Yugoslavia). In an analysis using multivariate logistic regression, the following factors were significantly related to prostate cancer: (1) occupational physical activity during the year preceding the disease [odds ratio (OR) = 3.87, 95% confidence interval (95% CI) = 2.09 - 7.16]; (2) occupational exposure to asbestos, steel, dyes and lacquers, bitumen, pitch, iron, nickel, lead, fertilizer and certain other agents (OR = 2.13, 95% CI = 1.05 - 4.32); (3) nephrolithiasis (OR = 4.52, 95% CI = 1.43 - 15.30); (4) 'other' diseases in medical history such as chronic bronchitis, chronic rheumatic diseases, hypertension, cardiomyopathy, diabetes mellitus, renal diseases, eye diseases and tuberculosis (OR = 3.14, 95% CI = 1.56 - 6.33); (5) a greater number (≤ 3) of brothers [OR = 2.08, 95% CI = 1.35 - 3.22); (6) greater numbers (≤ 8) of sexual partners (OR = 2.24, 95% CI = 1.13 - 4.44). Marital status, age at first marriage, educational level, age at first sexual intercourse, frequency of sexual intercourse, venereal diseases, tonsillectomy, appendectomy, hernia inguinale and hydrocele, anthropometric characteristics, smoking history, sport and recreational activities and family history of prostatic neoplasms were not found to be independently related to prostate cancer.

One hundred and one patients with histologically confirmed prostate cancer and 202 hospital controls individually matched by age (± 2 years), hospital admittance and place of residence, were interviewed during the period 1990-94 in two towns in central Serbia (Yugoslavia). In an analysis using multivariate logistic regression, the following factors were significantly related to prostate cancer: (1) occupational physical activity during the year preceding the disease [odds ratio (OR) = 3.87, 95% confidence interval (95% CI) = 2.09 - 7.16]; (2) occupational exposure to asbestos, steel, dyes and lacquers, bitumen, pitch, iron, nickel, lead, fertilizer and certain other agents (OR = 2.13, 95% CI = 1.05 - 4.32); (3) nephrolithiasis (OR = 4.52, 95% CI = 1.43 - 15.30); (4) 'other' diseases in medical history such as chronic bronchitis, chronic rheumatic diseases, hypertension, cardiomyopathy, diabetes mellitus, renal diseases, eye diseases and tuberculosis (OR = 3.14, 95% CI = 1.56 - 6.33); (5) a greater number (≤ 3) of brothers [OR = 2.08, 95% CI = 1.35 - 3.22); (6) greater numbers (≤ 8) of sexual partners (OR = 2.24, 95% CI = 1.13 - 4.44). Marital status, age at first marriage, educational level, age at first sexual intercourse, frequency of sexual intercourse, venereal diseases, tonsillectomy, appendectomy, hernia inguinale and hydrocele, anthropometric characteristics, smoking history, sport and recreational activities and family history of prostatic neoplasms were not found to be independently related to prostate cancer.

Objective: To compare clinical characteristics of the involuntary movements in primary and symptomatic dystonias. Patients and methods: 132 consecutive patients with the diagnosis of primary dystonia and 51 consecutive patients with secondary dystonia caused by well defined structural lesion(s) of the central nervous system, with particular emphasis on the characteristics of involuntary movements. Results: Eight variables with the highest risk contribution to either symptomatic or primary dystonias were identified: dystonic movement in secondary dystonia was much more frequently presented at rest, whereas the presence of dystonic tremor, chronic inflammatory process, or peripheral trauma located in the region that is later affected by dystonia, as well as the use of sensory tricks and development of spontaneous remissions, classified the affected patients more often in the category of those with primary dystonia. Conclusion: The study identified several clinical features that may be helpful in differentiating primary from secondary dystonia.

Objective: The study aimed to correlate the status of hepatitis C (HCV) and hepatitis B virus (HBV) co-infection in patients with human immunodeficiency virus (HIV) infection with clinical and demographic data prior to starting highly active antiretroviral therapy (HAART) and assess the impact of HCV and HBV co-infection on the natural history of HIV infection. Patients and Methods: The study involved a total of 836 treatment-naive patients with available serological status for HBV and HCV at the point of therapy initiation. Patients were stratified into four groups: HIV mono-infection, HIV/HCV, HIV/HBV, and HIV/HCV/HBV co-infection. Demographic, epidemiological, immunological and clinical characteristics were analyzed in order to assess the possible impact of HCV and HBV co-infection on HIV-related immunodeficiency and progression to AIDS. Results: The prevalence of HCV and HBV co-infection in our cohort was 25.7% and 6.3%, respectively. Triple HIV/HCV/HBV infection was recorded in 1.7% of the patients. In comparison with those co-infected with HCV, patients with HIV mono-infection had lower levels of serum liver enzymes activity and higher CD4 cell counts, and were less likely to have CD4 cell counts below100 cells/µL and clinical AIDS, with OR 0.556 and 0.561, respectively. No difference in the development of advanced immunodeficiency and/or AIDS was recorded between patients with HIV mono-infection and those co-infected with HBV, or both HCV/HBV. Conclusion: HIV/HCV co-infection was found to be more prevalent than HIV/HBV co-infection in a Serbian cohort. Co-infection with HCV was related to more profound immunodeficiency prior to therapy initiation, reflecting a possible unfavorable impact of HCV on the natural history of HIV infection. © 2018 Bentham Science Publishers.

To investigate possible relationships between diet and risk for bladder cancer in Serbia, the hospital-based case-control study was carried out. This study included 130 newly diagnosed bladder cancer patients and the same number of controls matched by sex, age (±2 years) and type of residence (rural or urban). Dietary information was obtained by using a food frequency questionnaire. Initial case-control comparisons were based on tertiles of average daily intake of control group. The odds ratios (ORs) were computed for each tertile, with the lowest tertile defined as the referent category. All variables (food items) significantly related to bladder cancer were included in multivariable logistic regression analysis. According to this analysis, risk factors for bladder cancer appeared to be consumption of liver (OR=6.60, 95%CI=1.89-23.03), eggs (OR=3.12, 95%CI=1.10-8.80), pork (OR=2.99, 95%CI=1.16-7.72), and pickled vegetable (OR=3.25, 95%CI=1.36-7.71). A protective effect was found for dietary intake of kale (OR=0.21, 95%CI=0.06-0.73), cereals (OR=0.19, 95%CI=0.06-0.62), tangerines (OR=0.21, 95%CI=0.07-0.68), cabbage (OR=0.27, 95% CI=0.11-0.68), and carrots (OR=0.15, 95%CI=0.05-0.41). The study indicated a potentially important role for dietary fat and pickled vegetables in bladder carcinogenesis. An inverse association was recorded between consumption of fruits, vegetables and cereals, and the development of bladder cancer. © Springer 2005.

Contemporary health-education intervention programs are increasingly used as a tool for improving health of school children [1-4]. Since 1992 a Network of 13 elementary Health Promoting Schools established in Yugoslavia (though not yet admitted to the European Network) has been operational. The Project was approved by the Ministries of Health, Education and Ecology from the very beginning, and financially supported by the Government of Serbia since 1995. The team of up to 40 health professionals, school principals and school project managers worked together for four years to change the working conditions in schools knowledge, attitudes and behaviour of school children and staff in order to improve their health [5]. The goal of this paper is to present results of health education intervention in changing of hygienic conditions in schools, as well as some of the attitudes, behaviour and knowledge of pupils and their parents. METHOD: The study took place before and after the intervention--at two points in time--during 1993 and 1996. The tri-angular approach including (1) pupils, (2) schools (teachers, school environment), and (3) parents, was used. A random pretest and post-test study design with control group (12 experimental and two control schools) has been implemented. The multiphase cluster sample was employed in order to represent all of the country typical regions. Six types of especially designed questionnaires were used to provide comparable variables in the sample of pupils, their parents and teachers. Exception were 1st-graders and 4th-graders for whom information were gathered by means of a "draw-and-write" investigation technique [6]. The response rate was 88.70% before and 98.28% after intervention. Data were analyzed by descriptive and inferential statistics using SPSS/PC software. RESULTS: Schools are somewhat less overcrowded, much cleaner and better maintained after the four-year intervention. Toilets are in a better condition, but there is still much more to be done. The 1st- and 4th-graders, surveyed by the draw-and-write method, mentioned numerous ways of keeping and improving health, which were summarized as fifteen health-improvement measures (Graph 1). The most frequently mentioned measures were nutrition, physical activity, general hygiene, oral and dental hygiene, sleeping and fresh air. Each of these measures was mentioned by over 20% of the surveyed pupils. Eleven of 15 variables showed significant increase in frequencies (at the level of at least p < 0.01) after the intervention. As an indicator of a nutrition behaviour, the regularity of main meals is analyzed (Graph 2). The majority of children eat regularly and the difference before and after intervention is significant only for the school meal (c2 = 30.715, p < 0.001). Although over 70% of children learn that general hygiene is good for health in junior graders, only about little more than 30% of senior graders have a bath or shower every day, while others only once or twice a week. The differences are significant before and after the intervention (c2 = 10.648, p < 0.05) only for everyday practice. More than 90% of senior-grade pupils brush their teeth at least once or several times a day. Over 60% of children in our survey go in for sport, whereas about 20% never do so. It seems that the intervention contributed symbolically to this practice important for health, though before-after difference is significant (c2 = 6.673, p < 0.05). However, in the control group schools children have much less physical activity in 1996, and this difference is significant (c2 = 14.070, p < 0.010). The psycho-emotional status of Yugoslav young people is strongly influenced by the situation in the country the war, the economic disaster and the impact of international sanctions. Consequently, more than one-fourth of the children complained of frequent exhaustion and concentration problems, which their parents also noticed. (ABSTRACT TRUNCATED)

The aim of the study was to evaluate association of type 1 diabetes in children and adolescents with positive family history of type 1 diabetes, type 2 diabetes, and thyroid, adrenal, rheumatic, allergic, celiac and some other diseases. A case-control study was conducted in Belgrade. The case group comprised 105 subjects ≤16 years old who were for the first time hospitalized because of type 1 diabetes during the period 1994-1997. For each case, two controls were chosen among children and adolescents treated for skin diseases. Cases and controls were individually matched by age (± one year), sex and place of residence (all were from Belgrade). In the statistical analysis we used X2 test, Fisher's exact test and univariate and multivariate logistic regressions. According to multivariate logistic regression analysis, risk of type 1 diabetes was significantly associated with a positive family history for type 1 diabetes (OR=8.58; 95% CI, 3.28-22.46), type 2 diabetes (OR=4.04; 95% CI, 2.31-7.07), allergic diseases (OR=3.32; 95% CI, 1.63-6.76), celiac and Crohn's diseases (OR=11.02; 95% CI, 1.14-106.89) and other diseases (thrombocytopenia, alopecia areata, psoriasis, chronic uveitis and pernicious anemia; OR=3.63; 95% CI, 1.05-12.48).

The aim of the study was to evaluate association of type 1 diabetes in children and adolescents with positive family history of type 1 diabetes, type 2 diabetes, and thyroid, adrenal, rheumatic, allergic, celiac and some other diseases. A case-control study was conducted in Belgrade. The case group comprised 105 subjects ≤16 years old who were for the first time hospitalized because of type 1 diabetes during the period 1994-1997. For each case, two controls were chosen among children and adolescents treated for skin diseases. Cases and controls were individually matched by age (± one year), sex and place of residence (all were from Belgrade). In the statistical analysis we used X2 test, Fisher's exact test and univariate and multivariate logistic regressions. According to multivariate logistic regression analysis, risk of type 1 diabetes was significantly associated with a positive family history for type 1 diabetes (OR=8.58; 95% CI, 3.28-22.46), type 2 diabetes (OR=4.04; 95% CI, 2.31-7.07), allergic diseases (OR=3.32; 95% CI, 1.63-6.76), celiac and Crohn's diseases (OR=11.02; 95% CI, 1.14-106.89) and other diseases (thrombocytopenia, alopecia areata, psoriasis, chronic uveitis and pernicious anemia; OR=3.63; 95% CI, 1.05-12.48).

Conditional logistic regression analysis of case-control study data showed that the following factors were significantly related to rheumatic fever occurrence: home dampness, change of place of residence during the last 5 years, low education of mother, body weight below normal, frequent sore throat and positive family history of rheumatic fever. © 1991 Gustav Fischer.

Marketing principles in health care delivery systems, reducing financial funds for health institutions and permanently increasing costs of medical equipment are responsible for the growing influence of management on physician's autonomy and working satisfaction. A questionnaire research was carried out in order to determine the best method of examination of this influence in Belgrade. All tested persons were medical doctors working in a primary health centers for 7 hours per day, who examined about 26 patients every day. More than a half of physician improved their knowledge by reading professional publications 6 hours per week. In spite of the fact that they all worked in the same institution, they expressed different opinions regarding the organisation. In the physicians' opinion clinical freedom and working satisfaction were at the high level, and partly connected with management practice. At the same time they were very unsatisfactory with personal income and budgetary distribution in the institution. It is evident that the method described in this article is useful in the analysis of the present situation related to management and physician's practice in the country.

Influence of factors which fluctuate in time is usually described in term of residuals present in general linear or nonlinear models. The aim of this paper has been to identify relationship between mortality rates and major social and economic factors and to compare three methods in detecting (hidden) relations between them. Data on frequencies for all causes of death, e.g. mortality rates have been collected for Central Serbia for 21 years, from 1973. until 1993. We have used two models based on the general linear model, variance-covariance regression and correlation and Box-Jenkins time series method and one nonlinear model based on neural networks. The relationship has been established, according to our results. Among presented methods we consider the analysis based on multilayer neural networks techniques has several advantages over classical approach. © The authors 1996.

Influence of factors which fluctuate in time is usually described in term of residuals present in general linear or nonlinear models. The aim of this paper has been to identify relationship between mortality rates and major social and economic factors and to compare three methods in detecting (hidden) relations between them. Data on frequencies for all causes of death, e.g. mortality rates have been collected for Central Serbia for 21 years, from 1973. until 1993. We have used two models based on the general linear model, variance-covariance regression and correlation and Box-Jenkins time series method and one nonlinear model based on neural networks. The relationship has been established, according to our results. Among presented methods we consider the analysis based on multilayer neural networks techniques has several advantages over classical approach. © The authors 1996.

Prevention of AIDS is very complex and requires, apart from basic knowledge about modes of transmission and protection, individual awareness of personal risk of infection and modification of one's risky behavior. Perception of risk has been suggested as an important element of sexual behavior change among people who engage in behaviors that place them at risk of HIV infection. This study investigates perceived risk of an HIV infection among inhabitants of municipality of Palilula. The purpose of this study was to determine the relative accuracy of risk assessment, especially for those with high-risk sexual behavior and to assess the factors related to self perception of developing AIDS. The survey was conducted as a household survey in Belgrade municipality "Palilula". The design of the survey was based on a two-stage stratified random sample. The population of Palilula municipality was divided in two stratums and 840 households were selected from each stratum. All members of selected households aged 15-49 years were interviewed. Total of 2659 persons was interviewed. Self administered questionnaire, developed by WHO and described in the Evaluation of National AIDS Programme: A Methods Package 1. Prevention of HIV infection, Geneva, 1994, was translated and adapted for local application. For analysis of findings methods of descriptive statistics, chi square test and logistic regression were used. Statistical analysis was performed using SPSS/PC. The results showed that most of the respondents (1662, 62.5%) do not perceive themselves to be at risk of HIV infection. Interestingly, almost one quarter of the sample couldn't estimate personal chances of being HIV infected. Significantly larger proportion of respondents living in urban area than those living in rural area perceived themselves at risk (17.1% of urban vs. 7.7% from rural area, chi square = 54.06, Ir0.00). Significant differences between age groups were found for perception of personal risk. Respondents up to 19 and those older than 25 in more than 60 percent and only a half of those between 20 and 24 believed that there was no chance they could become HIV infected. Almost 16% of respondents between 20 and 24 believed that they had some chances of contracting HIV, also approximately one third of them couldn't estimate personal risk of contracting HIV infection (chi square = 2.10, p = 0.00)(Graphe 1). Those who engaged in sex with irregular partners were found to perceive themselves in higher risk then those who were engaged in sex only with regular partners in previous 12 months. One fifth of those with irregular partners estimated their chances as good versus only 10% of those with only regular partners. A logistic regression analysis found that those who believed they were personally at risk knew someone with HIV, had more irregular sexual partners in last 12 months, were younger, and lived in urban area (Table 1). People correctly perceived the risk of HIV infection as being minimal as most of them didn't engage in risk behavior. Risk perception was higher among those who had more irregular partners, younger and those from urban area. Most important predictor of risk perception was knowledge of HIV positive individual.

Aims. This study sought to determine predictors of left ventricular thrombus formation and resolution after acute anterior wall myocardial infarction. Methods and Results. We have analysed clinical, echo-cardiographic and angiographic data in 53 consecutive patients with anterior myocardial infarction. Two dimensional and Doppler echocardiographic examinations were performed on days 1, 2, 3 and 7, after 3 and 6 weeks, and 3, 6, and 12 months following infarction. Coronary angiography was performed in 44 patients before hospital discharge. Left ventricular thrombus was detected in 30/53 patients (29/30 in the first week after infarction). Univariate analysis showed that left ventricular thrombus formation was associated with a higher initial end-systolic volume index (beta = 0.04, P = 0.001), and end-diastolic volume index (beta = 0.03, P = 0.03), a larger infarct perimeter (beta = 0.02, P = 0.01), a lower initial ejection fraction (beta = -0.06, P = 0.001), a higher initial wall motion score index (beta = 1.75, P = 0.023), a higher peak creatine kinase level (beta = 3.90, P = 0 .01), Killip class > 1 (beta = 1.11. P = 0.003), infarct expansion (beta = 0.78, P = 0.03), occluded infarct-related artery (beta = -0.87, P = 0.04) and non-thrombolytic therapy (beta= -0.76, P = 0.047). According to the Cox proportional regression model, independent predictors of thrombus formation after anterior myocardial infarction were high end-systolic volume index (beta = 0.06 , P = 0.01) and high peak creatine kinase level (beta = 5.17, P = 0.046). Thrombus disappeared in 11/30 (36.7%) patients during one-year echo-cardiographic follow-up. The only independent predictor of thrombus disappearance after acute myocardial infarction was the absence of apical dyskinesis 6 weeks after infarction (beta = -1.53, P = 0.045). Conclusions. Our data demonstrate that the best predictor of left ventricular thrombus formation after acute anterior myocardial infarction is a high initial end-systolic volume. Thrombus resolution is more likely to occur in patients without apical dyskinesis at the end of the healing phase of infarction.

In a hospital-based study of 119 patients with definite multiple sclerosis, demographic and clinical factors were analysed with respect to their validity in assessing the long-term prognosis. Over a mean follow-up of 21.7 years, the following factors negatively influenced the prognosis by the univariate analysis: male sex, age at onset over 25, pyramidal involvement or spasticity at onset, ≥ 3 functional systems affected at onset or after 5 years, incomplete first remission, length of the first remission ≤ 1 year, > 5 attacks in the first 10 years, secondary or primary-progressive disease, time to reach secondary progression over 5 years and time to reach EDSS 6 over 7 years. The multivariate model showed that in patients with relapsing-remitting disease, 5 years after onset, pyramidal involvement at onset and shorter time to reach EDSS 6 predicted poor outcome, while after 10 years, higher age at onset and incomplete first remission indicated poor prognosis. Ten years after onset, the predictors of poor outcome in the secondary-progressive group were shorter time to reach EDSS 6 or secondary progression and higher EDSS, while in the primary-progressive group those variables were spasticity or higher number of functional systems affected at onset, and higher EDSS after 5 and 10 years.

In a hospital-based study of 119 patients with definite multiple sclerosis, demographic and clinical factors were analysed with respect to their validity in assessing the long-term prognosis. Over a mean follow-up of 21.7 years, the following factors negatively influenced the prognosis by the univariate analysis: male sex, age at onset over 25, pyramidal involvement or spasticity at onset, ≥ 3 functional systems affected at onset or after 5 years, incomplete first remission, length of the first remission ≤ 1 year, > 5 attacks in the first 10 years, secondary or primary-progressive disease, time to reach secondary progression over 5 years and time to reach EDSS 6 over 7 years. The multivariate model showed that in patients with relapsing-remitting disease, 5 years after onset, pyramidal involvement at onset and shorter time to reach EDSS 6 predicted poor outcome, while after 10 years, higher age at onset and incomplete first remission indicated poor prognosis. Ten years after onset, the predictors of poor outcome in the secondary-progressive group were shorter time to reach EDSS 6 or secondary progression and higher EDSS, while in the primary-progressive group those variables were spasticity or higher number of functional systems affected at onset, and higher EDSS after 5 and 10 years.

With the proliferation of potential prognostic factors for lung cancer, it is becoming increasingly more difficult to integrate the information provided by these factors into a single accurate prediction of clinical outcome. Here we reviewed five classification methods for their capabilities in classification of 200 patients with lung cancer into distinct prognostic groups using survival outcome as a criteria. The source of patient data for this study is a Lung Tumour Registry from Institute for Lung Diseases, University Clinical Hospital, Belgrade. Almost all developed classification algorithms determined prognostic groups according to biochemical tumour markers LDH and alkaline phosphatase, producing most significant split, instead of commonly used staging variables. The choice of which approach to use for a given classification problem depends not only on statistical properties of method, but also on medical considerations, such as whether more differential findings are given greater weight and the applicability of a classification rule. © 1997, The authors.