Yan Cheng

Background: Individuals with heart failure (HF) are at increased risk of magnesium deficiency. Magnesium supplements are widely available and being used without clear evidence of efficacy in HF. Methods: We emulated a target trial to assess the association between magnesium supplements and adverse outcomes in U.S. veterans with newly diagnosed HF. Eligible patients were outpatients who received ambulatory care in the Veterans Health Administration between 1 January 2000 and 31 December 2020. Veterans with a hospitalization within 30 days prior to the eligible date, previous magnesium supplement or replacement use, or end-stage renal disease were ineligible for the trial. Initial self-reported magnesium supplement use (measured at eligible date) was identified in medical records using natural language processing and then checked repeatedly to confirm continuous use. The outcome was all-cause hospitalization or death. Patients were followed for up to five years from the eligible date and were censored if they changed from the assigned treatment strategy or initiated prescribed magnesium replacement. We applied inverse probability treatment weighting and Cox’s regression to estimate hazard ratios (HRs), with sensitivity analyses in patients surviving ≥ 6 months and those with continuous documentation of magnesium supplement use. Results: We enrolled 9900 magnesium supplement users and 9900 matched non-users. In the weighted cohort (mean age 72.6 years; 12.6% African American; 3.4% women; median follow-up 0.7 years), users had significantly better survival in both primary and sensitivity analyses (HR in primary analysis: 0.81 [0.77–0.86], p < 0.0001; HRs in sensitivity analyses: 0.91 [0.85–0.97], p = 0.0025 and 0.77 [0.72–0.82], p < 0.0001, respectively). Conclusions: magnesium supplement use was associated with a reduced risk of all-cause mortality or hospitalization among veterans with HF.

Background: Both diabetes and low magnesium‐containing food intake may increase the risk of heart failure (HF). However, the effect of nonprescription magnesium supplements on the risk of HF or major adverse cardiac events in patients with diabetes is unknown. Methods and Results: Using a target‐trial‐emulation approach, we assembled a national cohort of 94 239 veterans ≥40 years with diabetes, without prior HF or magnesium use, who received ambulatory care in the US veterans‐health care system documented by electronic clinic notes between January 1, 2006 and December 31, 2020. A natural language processing approach was used to detect self‐reported magnesium‐supplement use from clinic notes, n=17 619 were identified as users versus n=76 620 as nonusers. Using inverse probability treatment weighting, we constructed a cohort balanced in 88 baseline characteristics between users and nonusers. The primary outcome was incident HF. Secondary outcomes were major adverse cardiac events (myocardial infarction, stroke, HF hospitalization, or death). Hazard ratios (HRs) associated with magnesium‐supplement use and outcomes were estimated in the inverse probability treatment weighting weighted cohort using Cox regression. The inverse probability treatment weighting weighted cohort had a mean age of 67.4±10.3 years; 18.4% were Black, and 5.1% were women. The mean duration of magnesium‐supplement use was 3.5±3.1 (interquartile range, 1.1–5.1) years. Incident HF occurred in 8.0% of users and 9.7% of nonusers of magnesium supplements (HR, 0.94 [95% CI, 0.89–0.99]). Magnesium‐supplement use was also associated with a reduced risk of major adverse cardiac events (HR, 0.94 [95% CI, 0.90–0.97]). Conclusions: Long‐term nonprescription magnesium supplement use was associated with a lower risk of incident HF and major adverse cardiac events in patients with diabetes. These findings should be replicated in randomized controlled trials.

Background: Prior research on immigration and cognitive functioning was limited with conflicting findings. This study seeks to understand the association between immigration status and cognitive function in a representative sample of the US population and examine physical activity and nutritional intake’s interaction with immigration status.Method: The study cohort included 2,898 participants (673 immigrants and 2225 non‐immigrants) from US National Health and Nutrition Examination Survey (NHANES) 2011‐2014 who underwent at least one cognitive function test. Demographics, vital signs, cardiovascular diseases, physical activity, and nutrition were compared between immigrants and non‐immigrants by mean or proportion. Standardized z‐scores were calculated for each nutrient to ensure comparability. Cognitive function was assessed through four test scores (animal fluency, digit symbol, Consortium to Establish a Registry for Alzheimer’s Disease [CERAD], and global score). To account for potential cognitive disparities, standardized z‐scores were calculated for cognitive function scores within age, education, and language strata. Scores in the lowest quartile were classified as low. Multivariate logistic regressions were conducted to estimate the association between immigration status and risk of low standardized cognitive function scores, adjusting for all covariates. NHANES survey weights were applied to ensure the results are representative of US civilians.Result: Compared to the non‐immigrants, the immigrants in our sample were younger and less educated. (Table 1). Our analysis found that being an immigrant was significantly associated with a low global cognitive function score (OR = 2.13, 95% CI = (1.34, 3.40), p = 0.0024) (Table 2). Higher protein, folate and theobromine intakes were associated with a higher global cognitive function score, while higher carbohydrate and zinc intakes were associated with a lower score. No interaction was found between nutrient intakes and immigration status. Increased physical activity was associated with the reduced risk of low animal fluency score, especially beneficial in non‐immigrants (Figure 1).Conclusion: Being an immigrant was associated with low cognitive function after controlling age, education, language, and other covariates. A higher intake of protein, folate, and theobromine were found to be protective, though they did not have significant interactions with immigrant status.

Opioid use disorder is known to be under-coded as a diagnosis, yet problematic opioid use can be documented in clinical notes, which are included in electronic health records. We sought to identify problematic opioid use from a full range of clinical notes and compare the demographic and clinical characteristics of patients identified as having problematic opioid use exclusively in clinical notes to patients documented through ICD opioid use disorder diagnostic codes. We developed and applied a natural language processing (NLP) tool that combines rule-based pattern analysis and a trained support vector machine to the clinical notes of a patient cohort (n = 222,371) from two Veteran Affairs service regions to identify patients with problematic opioid use. We also used a set of ICD diagnostic codes to identify patients with opioid use disorder from the same cohort. The NLP tool achieved 96.6% specificity, 90.4% precision/PPV, 88.4% sensitivity/recall, and 94.4% accuracy on unseen test data. NLP exclusively identified 57,331 patients; 6997 patients had positive ICD code identifications. Patients exclusively identified through NLP were more likely to be women. Those identified through ICD codes were more likely to be male, younger, have concurrent benzodiazepine prescriptions, more comorbidities, and more care encounters, and were less likely to be married. Patients in both these groups had substantially elevated comorbidity levels compared with patients not documented through either method as experiencing problematic opioid use. Clinicians may be reluctant to code for opioid use disorder. It is therefore incumbent on the healthcare team to search for documentation of opioid concerns within clinical notes.

Mid-life high blood pressure (BP) is a risk factor for cerebral microinfarcts. Less is known about the relationship between late-life BP and cerebral microinfarcts, the examination of which is the objective of the current study. This case–control study analyzed data from 551 participants (94.6% aged ≥80 years; 58.6% women) in the Adult Changes in Thought (ACT) study who had autopsy data on microinfarcts and four values of systolic and diastolic blood pressure (SBP and DBP) before death. Using the average of four values, SBP was categorized using 10 mmHg intervals; a trend was defined as a ≥10 mmHg rise or fall from the first to fourth values (average gap of 6.5 years). Multivariable-adjusted regression models were used to examine the associations of BP and microinfarcts, adjusting for age, sex, last BP-to-death time, APOE genotype, and antihypertensive medication use. Microinfarcts were present in 274 (49.7%) participants; there were multiple in 51.8% of the participants, and they were located in cortical areas in 40.5%, subcortical areas in 29.6%, and both areas in 29.9% of the participants. All SBP categories (reference of 100–119 mmHg) and both SBP trends were associated with higher odds of both the presence and number of microinfarcts. The magnitude of these associations was numerically greater for subcortical than cortical microinfarcts. Similar associations were observed with DBP. These hypothesis-generating findings provide new information about the overall relationship between BP and cerebral microinfarcts in octogenarians.

Transformer is the latest deep neural network (DNN) architecture for sequence data learning, which has revolutionized the field of natural language processing. This success has motivated researchers to explore its application in the healthcare domain. Despite the similarities between longitudinal clinical data and natural language data, clinical data presents unique complexities that make adapting Transformer to this domain challenging. To address this issue, we have designed a new Transformer-based DNN architecture, referred to as Hybrid Value-Aware Transformer (HVAT), which can jointly learn from longitudinal and non-longitudinal clinical data. HVAT is unique in the ability to learn from the numerical values associated with clinical codes/concepts such as labs, and in the use of a flexible longitudinal data representation called clinical tokens. We have also trained a prototype HVAT model on a case-control dataset, achieving high performance in predicting Alzheimer’s disease and related dementias as the patient outcome. The results demonstrate the potential of HVAT for broader clinical data-learning tasks.

Cerebral microinfarcts are associated with cognitive impairment and dementia. Small vessel diseases such as cerebral arteriolosclerosis and cerebral amyloid angiography (CAA) have been found to be associated with microinfarcts. Less is known about the associations of these vasculopathies with the presence, numbers, and location of microinfarcts. These associations were examined in the clinical and autopsy data of 842 participants in the Adult Changes in Thought (ACT) study. Both vasculopathies were categorized by severity (none, mild, moderate, and severe) and region (cortical and subcortical). Odds ratios (OR) and 95% CIs for microinfarcts associated with arteriolosclerosis and CAA adjusted for possible modifying covariates such as age at death, sex, blood pressure, APOE genotype, Braak, and CERAD were estimated. 417 (49.5%) had microinfarcts (cortical, 301; subcortical, 249), 708 (84.1%) had cerebral arteriolosclerosis, 320 (38%) had CAA, and 284 (34%) had both. Ors (95% CI) for any microinfarct were 2.16 (1.46–3.18) and 4.63 (2.90–7.40) for those with moderate (n = 183) and severe (n = 124) arteriolosclerosis, respectively. Respective Ors (95% CI) for the number of microinfarcts were 2.25 (1.54–3.30) and 4.91 (3.18–7.60). Similar associations were observed for cortical and subcortical microinfarcts. Ors (95% Cis) for the number of microinfarcts associated with mild (n = 75), moderate (n = 73), and severe (n = 15) amyloid angiopathy were 0.95 (0.66–1.35), 1.04 (0.71–1.52), and 2.05 (0.94–4.45), respectively. Respective Ors (95% Cis) for cortical microinfarcts were 1.05 (0.71–1.56), 1.50 (0.99–2.27), and 1.69 (0.73–3.91). Respective Ors (95% Cis) for subcortical microinfarcts were 0.84 (0.55–1.28), 0.72 (0.46–1.14), and 0.92 (0.37–2.28). These findings suggest a significant association of cerebral arteriolosclerosis with the presence, number, and location (cortical and subcortical) of microinfarcts, and a weak and non-significant association of CAA with each microinfarct, highlighting the need for future research to better understand the role of small vessel diseases in the pathogenesis of cerebral microinfarcts.

There is widespread use of dietary supplements, some prescribed but many taken without a physician’s guidance. There are many potential interactions between supplements and both over-the-counter and prescription medications in ways that are unknown to patients. Structured medical records do not adequately document supplement use; however, unstructured clinical notes often contain extra information on supplements. We studied a group of 377 patients from three healthcare facilities and developed a natural language processing (NLP) tool to detect supplement use. Using surveys of these patients, we investigated the correlation between self-reported supplement use and NLP extractions from the clinical notes. Our model achieved an F1 score of 0.914 for detecting all supplements. Individual supplement detection had a variable correlation with survey responses, ranging from an F1 of 0.83 for calcium to an F1 of 0.39 for folic acid. Our study demonstrated good NLP performance while also finding that self-reported supplement use is not always consistent with the documented use in clinical records.

Alzheimer’s disease (AD) is characterized by cognitive impairment in the presence of cerebral amyloid plaques and neurofibrillary tangles. Less is known about the characteristics and predictors of resilience to cognitive impairment in the presence of neuropathological evidence of AD, the focus of this study. Of 3170 adults age ≥65 years in the National Alzheimer’s Coordinating Center (NACC) brain autopsy cohort, 1373 had evidence of CERAD level moderate to frequent neuritic plaque density and Braak stage V–VI neurofibrillary tangles. Resilience was defined by CDR-SOB and CDR-Global scores of 0–2.5 and 0–0.5, respectively, and non-resilience, CDR-SOB and CDR-Global scores >2.5 and >0.5, respectively. Multivariable logistic regression models were used to examine the independent associations of patient characteristics with resilience. There were 62 participants (4.8%) with resilience. Those with resilience were older (mean age, 88.3 vs. 82.4 years), more likely to be women (61.3% vs. 47.3%) and had a lower prevalence of the APOE-e4 carrier (41.9% vs. 56.2%). They also had a higher prevalence of hypertension, heart failure, atrial fibrillation, diuretic use, beta-blocker use, and APOE-e2 carrier status. Greater age at death, diuretic use, and APOE-e2 were the only characteristics independently associated with higher odds of the AD resilience phenotype (adjusted OR, 1.09; 95% CI, 1.05–1.13; p < 0.01; 2.00 (1.04–3.87), p = 0.04, 2.71 (1.31–5.64), p < 0.01, respectively). The phenotype of resilience to cognitive impairment is uncommon in older adults who have neuropathological evidence of AD.