Search

Background There is a global increase in isolation of nontuberculous mycobacteria (NTM). The aim of the study was to analyze longitudinal trends of NTM identification and pattern of antimicrobial susceptibility testing. Methods NTM recovery rates, distribution of NTM species identification, and antimicrobial susceptibility pattern of NTM at Pusan National University Yangsan Hospital between January 2016 and December 2020 were retrospectively analyzed. Results A total of 52,456 specimens from 21,264 patients were submitted for mycobacterial culture, of which 2,521 from 1,410 patients were NTM positive over five years (January 2016 to December 2020). NTM isolation showed an increasing trend from 2016 to 2020 (p

Abstract Background Acidosis frequently occurs in severe acute kidney injury (AKI), and continuous renal replacement therapy (CRRT) can control this pathologic condition. Nevertheless, acidosis may be aggravated; thus, monitoring is essential after starting CRRT. Herein, we addressed the longitudinal trajectory of acidosis on CRRT and its relationship with worse outcomes. Methods The latent growth mixture model was applied to classify the trajectories of pH during the first 24 hours and those of C-reactive protein (CRP) after 24 hours on CRRT due to AKI (n = 1815). Cox proportional hazard models were used to calculate hazard ratios of all-cause mortality after adjusting multiple variables or matching their propensity scores. Results The patients could be classified into 5 clusters, including the normally maintained groups (1st cluster, pH = 7.4; and 2nd cluster, pH = 7.3), recovering group (3rd cluster with pH values from 7.2 to 7.3), aggravating group (4th cluster with pH values from 7.3 to 7.2), and ill-being group (5th cluster, pH

Background: The purpose of this study is to investigate the pro-con of re-implementation by administrative areas and the difference in perception of community water fluoridation in implemented and non-implemented areas after the community water fluoridation in Korea was suspended. Through this, we intend to provide basic data that can help find ways to increase the support and interest of local residents.Methods: The 601 questionnaires collected through the survey and statistical analysis was conducted using SPSS Statistics 28.0. Results: As a result of analyzing the perception of the community water fluoridation according to the understanding of fluorine, the proportion of people who were not recognized by both fluorine and community water fluoridation was the highest (p＜0.05). As a result of the analysis of the pro-con of re-implementation of community water fluoridation, the approval was high. Among those who responded in favor, the place of re-implementation showed that ‘implementation nationwide’ was high. As for the reason for favor, it was found that it was possible to prevent dental caries disease. The reason for the objection was the lack of knowledge about fluoride. Conclusion: The results of the survey for the pro-con of the re-implementation of community water fluoridation showed a higher degree of ‘agree’ and showed that people in the area where community water fluoridation was not implemented showed higher interest in oral health prevention and management. Through this, not only oral education, but also correct information on the implementation method of community water fluoridation, the benefits of community water fluoridation, and the facts that were misunderstood in the past, as well as oral education, can be provided to raise interest in community water fluoridation. It is thought that the expected effect of the re-implementation of community water fluoridation can be obtained if such activities are carried out.

Background/Aims: Based on the Kyoto classification of gastritis, mucosal atrophy, endoscopic intestinal metaplasia, fold enlargement, nodularity, and diffuse redness may be associated with gastric cancer and Helicobacter pylori (H. pylori) infection. In this study, we investigated the association between Kyoto scores based on the aforementioned five variables and current H. pylori infection. Materials and Methods We reviewed medical records of consecutive patients who underwent endoscopic biopsies between January and June 2019. The study included 687 patients (370 and 317 patients with H. pylori-negative and -positive results, respectively). The Kyoto score was evaluated by the endoscopist who performed the test and was reconfirmed by another endoscopist. The total Kyoto score was analyzed using a receiver operating characteristic (ROC) curve for each score from 0 to 8. Multivariate analysis was used to determine the variables associated with H. pylori infection. Results: The maximum value of the Youden index (which reflects the ideal cut-off score of the Kyoto score on the ROC curve) was a Kyoto score of 2 points (Youden index 0.5905). Nodularity (OR 24.69, 95% CI 8.57~71.16, P

The Korean Society for Electrolyte and Blood Pressure Research, in collaboration with the Korean Society of Nephrology, has published a clinical practice guideline (CPG) document for hyponatremia treatment. The document is based on an extensive evidence-based review of the diagnosis, evaluation, and treatment of hyponatremia with the multidisciplinary participation of representative experts in hyponatremia with methodologist support for guideline development. This CPG consists of 12 recommendations (two for diagnosis, eight for treatment, and two for special situations) based on eight detailed topics and nine key questions. Each recommendation begins with statements graded by the strength of the recommendations and the quality of the evidence. Each statement is followed by rationale supporting the recommendations. The committee issued conditional recommendations in favor of rapid intermittent bolus administration of hypertonic saline in severe hyponatremia, the use of vasopressin receptor antagonists in heart failure with hypervolemic hyponatremia, and syndrome of inappropriate antidiuresis with moderate to severe hyponatremia, the individualization of desmopressin use, and strong recommendation on the administration of isotonic fluids as maintenance fluid therapy in hospitalized pediatric patients. We hope that this CPG will provide useful recommendations in practice, with the aim of providing clinical support for shared decision-making to improve patient outcomes.

Abstract Health-related quality of life (HRQOL) is an important issue among patients undergoing dialysis treatment. Peritoneal dialysis (PD) is associated with a number of adverse body composition changes. However, whether body composition is associated with HRQOL is uncertain. The purpose of this study was to analyze the effects of body composition on HRQOL in PD patients. We performed a cross-sectional observational study on the association between body composition and HRQOL in PD patients at a single center. Body composition was determined by multifrequency bioimpedance spectroscopy. HRQOL is summarized to three composite scores: kidney disease component summary (KDCS), physical component summary (PCS), and mental component summary (MCS). The relationships between HRQOL and the hydration index, lean tissue index (LTI), and fat tissue index (FTI) were analyzed by regression analysis. One hundred and ninety-seven PD patients were included in the present study. Patients with severe fluid overload showed a lower PCS. The hydration index and FTI showed statistically significant negative associations with PCS. In subgroup analysis, the associations between the hydration index and PCS remained robust after stratifying according to sex, age, and residual urine. Our results indicated that both the hydration index and FTI were negatively associated with HRQOL, especially PCS.

Abstract Glycemic variability (GV) is a risk factor for depression in patients with diabetes. However, whether it is also a predictor of incident depression in people without diabetes remains unclear. We aimed to investigate the association between visit-to-visit variability in fasting serum glucose (FSG) levels and the incidence of depression among Koreans without diabetes. This retrospective cohort study included data of people without diabetes who did not have depression at baseline and had at least three FSG measurements (n = 264,480) extracted from the 2002–2007 Korean National Health Insurance Service–National Health Screening Cohort. GV was calculated as the average successive variability of FSG. Among 264,480 participants, 198,267 were observed during 2008–2013 and their hazard ratios (HR) of incident depression were calculated. Participants with the highest GV showed a higher risk of depression in fully adjusted models than those with the lowest GV (HR, 1.09; 95% CI, 1.02–1.16). The risk of incident depression heightened with increasing GV (p for trend

Background/Aims Endoscopic channels are difficult to clean and can cause infection transmission. We examined the effectiveness of a newly developed channel-cleaning ball brush (BB), which is sucked into the endoscopic channel and scrapes and cleans the lumen as it passes through. Methods The upper and lower gastrointestinal endoscopes used for patient examinations were randomly selected as the conventional brush (CB) or BB group. After manual cleaning, the presence or absence of carbohydrates, proteins, adenosine triphosphate, and hemoglobin was assessed. Results Fifty-six and 58 endoscopes were cleaned with the CB and BB, respectively. Carbohydrate and protein were detected in one (1.8%) and two endoscopes (3.4%) in the CB and BB groups, respectively (p=1.000). Hemoglobin was observed in one (1.8%) and three endoscopes (5.2%) in the CB and BB groups, respectively (p=0.636). The adenosine triphosphate levels were 10.6±15.9 and 12.5±14.3 relative light units in the CB and BB groups, respectively (p=0.496). Twenty-seven (48.2%) and 19 (32.8%) endoscopes were positive for microbial cultures in the CB and BB groups, respectively (p=0.136). Conclusions The efficacy of BB was not significantly different from that of CB in the endoscopic channel-cleaning process.

Abstract Proteinuria is typically quantified according to the spot urine protein–creatinine ratio (UPCR) and an association with cardiovascular events has not been thoroughly investigated in chronic kidney disease (CKD) patients. We investigated whether the severity of proteinuria assessed by spot UPCR is associated with an increased risk for cardiovascular outcomes in the CKD population, and whether the relationship is influenced by urine creatinine concentration. We analyzed 1746 patients enrolled as part of The KoreaN cohort study for Outcome in patients With Chronic Kidney Disease (KNOW-CKD). Multivariable Cox proportional hazard analysis was performed to evaluate models with proteinuria as a predictor of renal events and extended major adverse cardiovascular events (eMACEs). Risk for renal events was significantly associated with proteinuria across all eGFR and UPCR categories. By contrast, risk for eMACEs increased significantly with UPCR in patients with eGFR ≥ 60 mL/min/1.73 m2 (hazard ratio [HR] 2.109; 95% confidence interval [CI] 1.375–3.235; P = 0.001), but not in patients with eGFR

Abstract Clinical guidelines recommend several risk factors to identify women in early pregnancy at high risk of developing pregnancy-associated hypertension. However, these variables result in low predictive accuracy. Here, we developed a prediction model for pregnancy-associated hypertension using graph-based semi-supervised learning. This is a secondary analysis of a prospective study of healthy pregnant women. To develop the prediction model, we compared the prediction performances across five machine learning methods (semi-supervised learning with both labeled and unlabeled data, semi-supervised learning with labeled data only, logistic regression, support vector machine, and random forest) using three different variable sets: [a] variables from clinical guidelines, [b] selected important variables from the feature selection, and [c] all routine variables. Additionally, the proposed prediction model was compared with placental growth factor, a predictive biomarker for pregnancy-associated hypertension. The study population consisted of 1404 women, including 1347 women with complete follow-up (labeled data) and 57 women with incomplete follow-up (unlabeled data). Among the 1347 with complete follow-up, 2.4% (33/1347) developed pregnancy-associated HTN. Graph-based semi-supervised learning using top 11 variables achieved the best average prediction performance (mean area under the curve (AUC) of 0.89 in training set and 0.81 in test set), with higher sensitivity (72.7% vs 45.5% in test set) and similar specificity (80.0% vs 80.5% in test set) compared to risk factors from clinical guidelines. In addition, our proposed model with graph-based SSL had a higher performance than that of placental growth factor for total study population (AUC, 0.71 vs. 0.80, p

The KoreaN Cohort Study for Outcomes in Patients With Chronic Kidney Disease (KNOW-CKD) was launched in 2011 with the support of the Korea Disease Control and Prevention Agency. The study was designed with the aim of exploring the various clinical features and characteristics of chronic kidney disease (CKD) in Koreans, and elucidating the risk factors for CKD progression and adverse outcomes of CKD. For the cohort study, nephrologists at 9 tertiary university-affiliated hospitals participated in patient recruitment and follow-up. Biostatisticians and epidemiologists also participated in the basic design and structuring of the study. From 2011 until 2016, the KNOW-CKD Phase I recruited 2238 adult patients with CKD from stages G1 to G5, who were not receiving renal replacement therapy. The KNOW-CKD Phase II recruitment was started in 2019, with an enrollment target of 1500 subjects, focused on diabetic nephropathy and hypertensive kidney diseases in patients with reduced kidney function who are presumed to be at a higher risk of adverse outcomes. As of 2021, the KNOW-CKD investigators have published articles in the fields of socioeconomics, quality of life, nutrition, physical activity, renal progression, cardiovascular disease and outcomes, anemia, mineral bone disease, serum and urine biomarkers, and international and inter-ethnic comparisons. The KNOW-CKD researchers will elaborate a prediction model for various outcomes of CKD such as the development of end-stage kidney disease, major adverse cardiovascular events, and death.

Background Mayo imaging classification (MIC) is a useful biomarker to predict disease progression in autosomal dominant polycystic kidney disease (ADPKD). This study was performed to validate MIC in the prediction of renal outcome in a prospective Korean ADPKD cohort and evaluate clinical parameters associated with rapid disease progression. Methods A total of 178 ADPKD patients were enrolled and prospectively observed for an average duration of 6.2 ± 1.9 years. Rapid progressor was defined as MIC 1C through 1E while slow progressor was defined as 1A through 1B. Renal composite outcome (doubling of serum creatinine, 50% decline of estimated glomerular filtration rate [eGFR], or initiation of renal replacement therapy) as well as the annual percent change of height-adjusted total kidney volume (mHTKV-α) and eGFR decline (mGFR-α) were compared between groups. Results A total of 110 patients (61.8%) were classified as rapid progressors. These patients were younger and showed a higher proportion of male patients. Rapid progressor was an independent predictor for renal outcome (hazard ratio, 4.09; 95% confidence interval, 1.23–13.54; p = 0.02). The mGFR-α was greater in rapid progressors (–3.58 mL/min per year in 1C, –3.7 in 1D, and –4.52 in 1E) compared with that in slow progressors (–1.54 in 1A and –2.06 in 1B). The mHTKV-α was faster in rapid progressors (5.3% per year in 1C, 9.4% in 1D, and 11.7% in 1E) compared with that in slow progressors (1.2% in 1A and 3.8% in 1B). Conclusion MIC is a good predictive tool to define rapid progressors in Korean ADPKD patients.

Abstract We aimed to compare the short- and long-term outcomes of simultaneous laparoscopic, hybrid, and open resection for colorectal cancer and synchronous liver metastases. We retrospectively analyzed the data of 647 patients with simultaneous resection of colorectal cancer and liver metastases between January 2006 and December 2018 at three tertiary referral hospitals. Patient’s baseline characteristics, perioperative outcomes, pathological examination results, liver-specific recurrence rate and survivals were compared between the propensity score-matched groups. Forty-two and 81 patients were selected for the laparoscopic vs. hybrid groups, and 48 and 136 patients for laparoscopic vs. open groups, respectively. The laparoscopic group had fewer wound complications (2.1 vs. 13.2%; p = 0.028) than the open group, and a shorter postoperative hospital stay than the hybrid and open groups (8 vs. 11 days, p

Background Metabolic syndrome (MetS) is reportedly a crucial risk factor for coronavirus disease 2019 (COVID-19). Since the epidemiological studies that examine this association are few and include small samples, we investigated the relationship between MetS and COVID-19 severity and death using a larger sample in the Republic of Korea. Methods We analyzed 66,321 patients, 4,066 of whom had COVID-19. We used chi-square tests to examine patients’ characteristics. We performed logistic regression analysis to analyze differences in COVID-19 infection and clinical outcomes according to the presence of MetS. Results Although MetS was not significantly associated with COVID-19 risk, acquiring MetS was significantly associated with the risk of severe COVID-19 outcomes (odds ratio [OR], 1.97; 95% confidence interval [CI], 1.34 to 2.91; P=0.001). The mortality risk was significantly higher in COVID-19 patients with MetS (OR, 1.74; 95% CI, 1.17 to 2.59; P=0.006). Patients with abnormal waist circumference were approximately 2.07 times more likely to develop severe COVID-19 (P

Background Appropriate monitoring of intradialytic biosignals is essential to minimize adverse outcomes because intradialytic hypotension and arrhythmia are associated with cardiovascular risk in hemodialysis patients. However, a continuous monitoring system for intradialytic biosignals has not yet been developed. Methods This study investigated a cloud system that hosted a prospective, open-source registry to monitor and collect intradialytic biosignals, which was named the CONTINUAL (Continuous mOnitoriNg viTal sIgN dUring hemodiALysis) registry. This registry was based on real-time multimodal data acquisition, such as blood pressure, heart rate, electrocardiogram, and photoplethysmogram results. Results We analyzed session information from this system for the initial 8 months, including data for some cases with hemodynamic complications such as intradialytic hypotension and arrhythmia. Conclusion This biosignal registry provides valuable data that can be applied to conduct epidemiological surveys on hemodynamic complications during hemodialysis and develop artificial intelligence models that predict biosignal changes which can improve patient outcomes.

Purpose Local excision (LE) is an alternative initial treatment for clinical T1 rectal cancer, and has avoided potential morbidity. This study aimed to evaluate the clinical outcomes of LE compared with total mesorectal excision (TME) for clinical T1 rectal cancer. Methods Between January 2000 and December 2011, we retrospectively reviewed from multicenter data in patients with clinically suspected T1 rectal cancer treated with either LE or TME. Of 1,071 patients, 106 were treated with LE and 965 were treated with TME. The data were analyzed using propensity score matching, with each group comprising 91 patients. Results After propensity score matching, the median follow-up time was 60.8 months (range, 0.6–150.6 months). After adjustment for the necessary variables, patients who underwent LE showed a significantly higher local recurrence rate than did those who underwent TME; however, there were no differences in disease-free survival and overall survival. In the multivariate analysis, age (hazard ratio [HR], 9.620; 95% confidence interval [CI], 3.415–27.098; P

Background/Aims: Bismuth-based quadruple therapy (BQT) is a treatment option for clarithromycin-resistant Helicobacter pylori (HP) infection. The aim of this study was to compare the efficacy of 7-day BQT with that of 14-day BQT as first-line treatment for clarithromycin-resistant HP infection. Methods: A total of 162 treatment-naïve patients with peptic ulcer disease and clarithromycinresistant HP infection confirmed by real-time polymerase chain reaction (RT-PCR) were enrolled. The enrolled patients were prospectively randomized to receive BQT for either 7 or 14 days of treatment. Eradication of HP infection was assessed using a 13C-urea breath test. Eradication and adverse event rates of the two groups were assessed. Results: The overall eradication rates in the intention-to-treat (ITT) and per-protocol (PP) analyses were 83.0% (95% confidence interval [CI], 77.2% to 88.9%; 132/159) and 89.8% (95% CI, 84.9% to 94.7%; 132/147), respectively. The eradication rates in the ITT analysis were 79.0% (64/81) in the 7-day group and 87.2% (68/78) in the 14-day group (p=0.170). The eradication rates in the PP analysis were 86.5% (64/74) in the 7-day group and 93.2% (68/73) in the 14-day group (p=0.182). Clinically significant adverse events occurred in 18.2% of patients. There was no statistically significant difference in the rates of individual or all adverse events between the two groups. Conclusions: Both 7-day and 14-day BQT were effective and safe as first-line therapy for HP infections identified as resistant to clarithromycin by RT-PCR. For clarithromycin-resistant HP infections, 7-day BQT may be sufficient as first-line therapy. [ABSTRACT FROM AUTHOR]

In this study, full-scale loading tests were conducted to investigate web-shear strengths of hollow-core slab (HCS) members strengthened in shear by using practically viable methods. All the HCS units used in the current test program were fabricated by using the individual mold method, not by the extrusion method, and the key experimental variables of the shear test were set as the presence of shear reinforcement, core-filling concrete, topping concrete, and also the magnitude of effective prestress. The shear force-displacement behaviors, crack patterns, and strain response of shear reinforcements were reported in detail. In addition, to identify the shear strength enhancement provided under various strengthening conditions in a quantitative manner, existing shear test results of series specimens, including a naked HCS member and corresponding composite HCS members with cast-in-place (CIP) concrete and/ or shear reinforcements, were collected from literature. On this basis, a practical design expression capable of estimating shear strengths of HCSs strengthened with CIP concrete and stirrups was suggested based on the ACI 318 code equation. The proposed method evaluated the shear strengths of the collected specimens with a good level of accuracy, regardless of the presence of core-filling concrete, topping concrete, and shear reinforcements. [ABSTRACT FROM AUTHOR]

Bio-derived materials could play a major role in future supportable green and health technologies. Nontraditional materials such as proteins and biopolymers are observed for their probable use in the fabrication of bioplastics for packaging. In the present work, novel ternary composite films of chitosan (CS), polyvinyl alcohol (PVA), with different ratios of natural egg white (EW) and glycerol as a plasticizer, were fabricated by solution casting technique. The synthesized ternary composite films were characterized by UV–Vis absorption spectra, Fourier transform-infrared spectroscopy (FT-IR), X-ray diffraction (XRD), and scanning electron microscopy (SEM) analysis. The hydrophobicity of as-synthesized samples was confirmed by contact angle measurement. The microbiological screening analysis was established against bacteria viz. Pseudomonas sp., Bacillus subtilis and Staphylococcus aureus bacteria. The attained results denoted that the film might be an appropriate applicant for food packaging applications. [ABSTRACT FROM AUTHOR]

Background Osteoprotegerin is an important regulator of bone metabolism and vascular calcification. The association between serum osteoprotegerin level and chronic kidney disease (CKD) progression has not been elucidated. We investigated the prognostic value of serum osteoprotegerin levels in nondialysis CKD patients. Methods We analyzed 2,082 patients enrolled in the Korean Cohort Study for Outcomes in Patients with CKD between 2011 and 2016. Patients were divided into quartiles by their serum osteoprotegerin levels. The primary outcome was the occurrence of ≥1 of the following: dialysis initiation, kidney transplantation, a two-fold increase in serum creatinine level from baseline, or a 50% decrease in the estimated glomerular filtration rate (eGFR). Cox proportional hazard regression models were used to investigate the prognostic value of the serum osteoprotegerin level to CKD progression. Results The median follow-up period was 48.9 months, and 641 patients (30.8%) experienced the primary outcome. The hazard ratio of serum osteoprotegerin for renal progression in the full extended Cox proportional hazard model was 1.064 (95% confidence interval, 1.041–1.088). Subgroup analyses by age, presence of diabetes, and eGFR showed significant results consistent with the overall analysis results. Conclusion Serum osteoprotegerin level is independently associated with renal prognosis and could have prognostic importance in CKD progression.

Nitrogen oxides (NOx; NO and NO2) are major air pollutants and can cause harmful effects on the human body. Electron Beam Flue Gas Treatment (EBFGT) is a technology that generates electrons with an energy of 0.5–1 MeV using electron accelerators and effectively processes exhaust gases. In this study, NOx was removed using an electron beam accelerator with spraying additives (NaOH and NH4OH). NO and NO2 were 100% and more than 94% removed, respectively, at an electron beam absorbed dose of 20 kGy and an additive concentration of 0.02 M (mol/L). In most cases, NOx was removed better with lower initial NOx concentrations and higher electron beam absorbed doses. As the irradiation strength (mA) of the electron beam increases, the probability of electron impact on the material accordingly rises, which may lead to increase removal efficiency. The results of the present study show that the continuous electron beam process using additives achieved more effective removal efficiency than either individual process (wet-scrubbing or EB irradiation only).

All-organic perovskites exhibit structural tunability and solution-processability, but are disadvantaged by a lower coercive field compared to inorganic ones. Here, the authors demonstrate that modulating hydrogen bond strength in such perovskites can generate a large coercive field.

Background Diabetic nephropathy (DN) can affect quality of life (QoL) because it requires arduous lifelong management. This study analyzed QoL differences between DN patients and patients with other chronic kidney diseases (CKDs). Methods The analysis included subjects (n = 1,766) from the KNOW-CKD (Korean Cohort Study for Outcomes in Patients with Chronic Kidney Disease) cohort who completed the Kidney Disease Quality of Life Short Form questionnaire. After implementing propensity score matching (PSM) using factors that affect the QoL of DN patients, QoL differences between DN and non-DN participants were examined. Results Among all DN patients (n = 390), higher QoL scores were found for taller subjects, and lower scores were found for those who were unemployed or unmarried, received Medical Aid, had lower economic status, had higher platelet counts or alkaline phosphatase levels, or used clopidogrel or insulin. After PSM, the 239 matched DN subjects reported significantly lower patient satisfaction (59.9 vs. 64.5, p = 0.02) and general health (35.3 vs. 39.1, p = 0.04) than the 239 non-DN subjects. Scores decreased in both groups during the 5-year follow-up, and the scores in the work status, sexual function, and role-physical domains were lower among DN patients than non-DN patients, though those differences were not statistically significant. Conclusion Socioeconomic factors of DN were strong risk factors for impaired QoL, as were high platelet, alkaline phosphatase, and clopidogrel and insulin use. Clinicians should keep in mind that the QoL of DN patients might decrease in some domains compared with non-DN CKDs.

Background/Aims To develop an early prediction model for gestational diabetes mellitus (GDM) using machine learning and to evaluate whether the inclusion of nonalcoholic fatty liver disease (NAFLD)-associated variables increases the performance of model. Methods This prospective cohort study evaluated pregnant women for NAFLD using ultrasound at 10–14 weeks and screened them for GDM at 24–28 weeks of gestation. The clinical variables before 14 weeks were used to develop prediction models for GDM (setting 1, conventional risk factors; setting 2, addition of new risk factors in recent guidelines; setting 3, addition of routine clinical variables; setting 4, addition of NALFD-associated variables, including the presence of NAFLD and laboratory results; and setting 5, top 11 variables identified from a stepwise variable selection method). The predictive models were constructed using machine learning methods, including logistic regression, random forest, support vector machine, and deep neural networks. Results Among 1,443 women, 86 (6.0%) were diagnosed with GDM. The highest performing prediction model among settings 1–4 was setting 4, which included both clinical and NAFLD-associated variables (area under the receiver operating characteristic curve [AUC] 0.563–0.697 in settings 1–3 vs. 0.740–0.781 in setting 4). Setting 5, with top 11 variables (which included NAFLD and hepatic steatosis index), showed similar predictive power to setting 4 (AUC 0.719–0.819 in setting 5, P=not significant between settings 4 and 5). Conclusions We developed an early prediction model for GDM using machine learning. The inclusion of NAFLD-associated variables significantly improved the performance of GDM prediction. (ClinicalTrials.gov Identifier: NCT02276144)