Ralf Rittner

Acute Myocardial Infarction Detected in the 12-Lead ECG by Artificial Neural Networks

BACKGROUND The 12-lead ECG, together with patient history and clinical findings, remains the most important method for early diagnosis of acute myocardial infarction. Automated interpretation of ECG is widely used as decision support for less experienced physicians. Recent reports have demonstrated that artificial neural networks can be used to improve selected aspects of conventional rule-based interpretation programs. The purpose of this study was to detect acute myocardial infarction in the 12-lead ECG with artificial neural networks. METHODS AND RESULTS A total of 1120 ECGs from patients with acute myocardial infarction and 10,452 control ECGs, recorded at an emergency department with computerized ECGs, were studied. Artificial neural networks were trained to detect acute myocardial infarction by use of measurements from the 12 ST-T segments of each ECG, together with the correct diagnosis. After this training process, the performance of the neural networks was compared with that of a widely used ECG interpretation program and the classification of an experienced cardiologist. The neural networks showed higher sensitivities and discriminant power than both the interpretation program and cardiologist. The sensitivity of the neural networks was 15.5% (95% confidence interval [CI], 12.4 to 18.6) higher than that of the interpretation program compared at a specificity of 95.4% (P<.00001) and 10.5% (95% CI, 7.2 to 13.6) higher than the cardiologist at a specificity of 86.3% (P<.00001). CONCLUSIONS Artificial neural networks can be used to improve automated ECG interpretation for acute myocardial infarction. The networks may be useful as decision support even for the experienced ECG readers.

Risk of musculoskeletal disorders among females and males in repetitive/constrained work

This paper combines epidemiological data on musculoskeletal morbidity in 40 female and 15 male occupational groups (questionnaire data 3720 females, 1241 males, physical examination data 1762 females, 915 males) in order to calculate risk for neck and upper limb disorders in repetitive/constrained vs. varied/mobile work and further to compare prevalence among office, industrial and non-office/non-industrial settings, as well as among jobs within these. Further, the paper aims to compare the risk of musculoskeletal disorders from repetitive/constrained work between females and males. Prevalence ratios (PR) for repetitive/constrained vs. varied/mobile work were in neck/shoulders: 12-month complaints females 1.2, males 1.1, diagnoses at the physical examination 2.3 and 2.3. In elbows/hands PRs for complaints were 1.7 and 1.6, for diagnoses 3.0 and 3.4. Tension neck syndrome, cervicalgia, shoulder tendonitis, acromioclavicular syndrome, medial epicondylitis and carpal tunnel syndrome showed PRs > 2. In neck/shoulders PRs were similar across office, industrial and non-office/non-industrial settings, in elbows/hands, especially among males, somewhat higher in industrial work. There was a heterogeneity within the different settings (estimated by bootstrapping), indicating higher PRs for some groups. As in most studies, musculoskeletal disorders were more prevalent among females than among males. Interestingly, though, the PRs for repetitive/constrained work vs. varied/mobile were for most measures approximately the same for both genders. In conclusion, repetitive/constrained work showed elevated risks when compared to varied/mobile work in all settings. Females and males showed similar risk elevations. This article enables comparison of risk of musculoskeletal disorders among many different occupations in industrial, office and other settings, when using standardised case definitions. It confirms that repetitive/constrained work is harmful not only in industrial but also in office and non-office/non-industrial settings. The reported data can be used for comparison with future studies.

Artificial neural networks for recognition of electrocardiographic lead reversal.

Misplacement of electrodes during the recording of an electrocardiogram (ECG) can cause an incorrect interpretation, misdiagnosis, and subsequent lack of proper treatment. The purpose of this study was twofold: (1) to develop artificial neural networks that yield peak sensitivity for the recognition of right/left arm lead reversal at a very high specificity; and (2) to compare the performances of the networks with those of 2 widely used rule-based interpretation programs. The study was based on 11,009 ECGs recorded in patients at an emergency department using computerized electrocardiographs. Each of the ECGs was used to computationally generate an ECG with right/left arm lead reversal. Neural networks were trained to detect ECGs with right/left arm lead reversal. Different networks and rule-based criteria were used depending on the presence or absence of P waves. The networks and the criteria all showed a very high specificity (99.87% to 100%). The neural networks performed better than the rule-based criteria, both when P waves were present (sensitivity 99.1%) or absent (sensitivity 94.5%). The corresponding sensitivities for the best criteria were 93.9% and 39.3%, respectively. An estimated 300 million ECGs are recorded annually in the world. The majority of these recordings are performed using computerized electrocardiographs, which include algorithms for detection of right/left arm lead reversals. In this study, neural networks performed better than conventional algorithms and the differences in sensitivity could result in 100,000 to 400,000 right/left arm lead reversals being detected by networks but not by conventional interpretation programs.

Maternal Exposure to Air Pollution and Birth Outcomes.

Background The knowledge about air pollution effects on birth weight, prematurity, and small for gestational age (SGA) in low-exposure areas is insufficient. Objectives The aim of this birth cohort study was to investigate whether low-level exposure to air pollution was associated with prematurity and fetal growth and whether there are sex-specific effects. Method We combined high-quality registry information on 81,110 births with individually modeled exposure data at residence for nitrogen oxides (NOx) and proximity to roads with differing traffic density. The data were analyzed by logistic and linear regression with and without potential confounders. Results We observed an increased risk for babies being SGA when we compared highest and lowest NOx quartiles, adjusting for maternal age, smoking, sex, and year of birth. After additional adjustment for maternal country of origin and parity (which were highly intercorrelated), the increase was no longer statistically significant. However, in subgroup analyses when we compared highest and lowest NOx quartiles we still observed an increased risk for SGA for girls [odds ratio (OR) = 1.12; 95% confidence interval (CI), 1.01–1.24); we also observed increased risk among mothers who had not changed residency during pregnancy (OR = 1.09; 95% CI, 1.01–1.18). The confounders with the greatest impact on SGA were parity and country of origin. Concerning prematurity, the prevalence was lower in the three higher NOx exposure quartiles compared with the lowest category. Conclusion For future studies on air pollution effects on birth outcomes, careful control of confounding is crucial.

Detection of frequently overlooked electrocardiographic lead reversals using artificial neural networks

Artificial neural networks can be used to recognize lead reversals in the 12-lead electrocardiogram at very high specificity, and the sensitivity is much higher than that of a conventional interpretation program. The neural networks developed in this and an earlier study for detection of lead reversals, in combination with an algorithm for the right arm/right foot lead reversal, would recognize approximately 75% of lead reversals encountered in clinical practice.

Population-based consultation patterns in patients with shoulder pain diagnoses.

BackgroundTo assess the annual consultation prevalence and new onset consultation rate for doctor-diagnosed shoulder pain conditions.MethodsWe identified all residents in the southernmost county in Sweden who received a shoulder pain diagnosis during 2006 (ICD-10 code M75). In subjects who did not consult due to such disorders during 2004 and 2005, we estimated the new onset consultation rate. The distribution of specific shoulder conditions and the length of the period of repeated consultation were calculated.ResultsAnnual consultation prevalence was 103/10 000 women and 98/10 000 men. New onset consultation rate was 80/10 000 women (peak in age 50–59 at 129/10 000) and 74/10 000 men (peak in age 60–69 at 116/10 000). About one fifth of both genders continued to consult more than three months after initial presentation, but only a few percent beyond two years. Rotator cuff - and impingement syndromes were the most frequent diagnoses.ConclusionThe annual consultation prevalence for shoulder pain conditions (1%) was similar in women and men, and about two thirds of patients consulted a doctor only once. Impingement and rotator cuff syndromes were the most frequent diagnoses.

Mesoamerican nephropathy: geographical distribution and time trends of chronic kidney disease mortality between 1970 and 2012 in Costa Rica

Objectives Mesoamerican nephropathy is an epidemic of chronic kidney disease (CKD) unrelated to traditional causes, mostly observed in sugarcane workers. We analysed CKD mortality in Costa Rica to explore when and where the epidemic emerged, sex and age patterns, and relationship with altitude, climate and sugarcane production. Methods SMRs for CKD deaths (1970–2012) among population aged ≥20 were computed for 7 provinces and 81 counties over 4 time periods. Time trends were assessed with age-standardised mortality rates. We qualitatively examined relations between mortality and data on altitude, climate and sugarcane production. Results During 1970–2012, age-adjusted mortality rates in the Guanacaste province increased among men from 4.4 to 38.5 per 100 000 vs 3.6–8.4 in the rest of Costa Rica, and among women from 2.3 to 10.7 per 100 000 vs 2.6–5.0 in the rest of Costa Rica. A significant moderate excess mortality was observed among men in Guanacaste already in the mid-1970s, steeply increasing thereafter; a similar female excess mortality appeared a decade later, remaining stable. Male age-specific rates were high in Guanacaste for age categories ≥30, and since the late 1990s also for age range 20–29. The male spatiotemporal patterns roughly followed sugarcane expansion in hot, dry lowlands with manual harvesting. Conclusions Excess CKD mortality occurs primarily in Guanacaste lowlands and was already present 4 decades ago. The increasing rates among Guanacaste men in hot, dry lowland counties with sugarcane are consistent with an occupational component. Stable moderate increases among women, and among men in counties without sugarcane, suggest coexisting environmental risk factors.

Measured and modeled personal and environmental NO2 exposure

BackgroundMeasured or modeled levels of outdoor air pollution are being used as proxies for individual exposure in a growing number of epidemiological studies. We studied the accuracy of such approaches, in comparison with measured individual levels, and also combined modeled levels for each subject’s workplace with the levels at their residence to investigate the influence of living and working in different places on individual exposure levels.MethodsA GIS-based dispersion model and an emissions database were used to model concentrations of NO2 at the subject’s residence. Modeled levels were then compared with measured levels of NO2. Personal exposure was also modeled based on levels of NO2 at the subject’s residence in combination with levels of NO2 at their workplace during working hours.ResultsThere was a good agreement between measured façade levels and modeled residential NO2 levels (rs = 0.8, p > 0.001); however, the agreement between measured and modeled outdoor levels and measured personal exposure was poor with overestimations at low levels and underestimation at high levels (rs = 0.5, p > 0.001 and rs = 0.4, p > 0.001) even when compensating for workplace location (rs = 0.4, p > 0.001).ConclusionModeling residential levels of NO2 proved to be a useful method of estimating façade concentrations. However, the agreement between outdoor levels (both modeled and measured) and personal exposure was, although significant, rather poor even when compensating for workplace location. These results indicate that personal exposure cannot be fully approximated by outdoor levels and that differences in personal activity patterns or household characteristics should be carefully considered when conducting exposure studies. This is an important finding that may help to correct substantial bias in epidemiological studies.

Fetal growth and air pollution - A study on ultrasound and birth measures

Abstract Air pollution has been suggested to affect fetal growth, but more data is needed to assess the timing of exposure effects by using ultrasound measures. It is also important to study effects in low exposure areas to assess eventual thresholds of effects. The MAPSS (Maternal Air Pollution in Southern Sweden) cohort consists of linked registry data for around 48,000 pregnancies from an ultrasound database, birth registry and exposure data based on residential addresses. Measures of air pollution exposure were obtained through dispersion modelling with input data from an emissions database (NOx) with high resolution (100–500 m grids). Air pollution effects were assessed with linear regressions for the following endpoints; biparietal diameter, femur length, abdominal diameter and estimated fetal weight measured in late pregnancy and birth weight and head circumference measured at birth. We estimated negative effects for NOx; in the adjusted analyses the decrease of abdominal diameter and femur length were −0.10 (−0.17, −0.03) and −0.13 (−0.17, −0.01) mm, respectively, per 10 &mgr;g/m3 increment of NOx. We also estimated an effect of NOx‐exposures on birth weight by reducing birth weight by 9 g per 10 &mgr;g/m3 increment of NOx. We estimated small but statistically significant effects of air pollution on late fetal and birth size and reduced fetal growth late in pregnancy in a geographic area with levels below current WHO air quality guidelines. HighlightsLargest study on ultrasound measures and air pollution to date.Individual data on potential confounders (also SES) and nitrogen oxides.Area with levels below current WHO air quality guidelines.Birth weight was reduced by 9 g per 10 &mgr;g/m3 increment of NOx.Small but statistically significant effects of air pollution on late fetal size.

Recognition of lead reversals in pediatric electrocardiograms

We have trained artificial neural networks to detect right/left arm lead reversals in pediatric electrocardiograms with a performance significantly higher than that of currently used methods. We believe that this type of neural network can be a valuable type of quality control in pediatric electrocardiography.