Multimodal fine-tuning of clinical language models for predicting COVID-19 outcomes

Clinical prediction models tend only to incorporate structured healthcare data, ignoring information recorded in other data modalities, including free-text clinical notes. Here, we demonstrate how multimodal models that effectively leverage both structured and unstructured data can be developed for predicting COVID-19 outcomes. The models are trained end-to-end using a technique we refer to as multimodal fine-tuning, whereby a pre-trained language model is updated based on both structured and unstructured data. The multimodal models are trained and evaluated using a multicenter cohort of COVID-19 patients encompassing all encounters at the emergency department of six hospitals. Experimental results show that multimodal models, leveraging the notion of multimodal fine-tuning and trained to predict (i) 30-day mortality, (ii) safe discharge and (iii) readmission, outperform unimodal models trained using only structured or unstructured healthcare data on all three outcomes. Sensitivity analyses are performed to better understand how well the multimodal models perform on different patient groups, while an ablation study is conducted to investigate the impact of different types of clinical notes on model performance. We argue that multimodal models that make effective use of routinely collected healthcare data to predict COVID-19 outcomes may facilitate patient management and contribute to the effective use of limited healthcare resources.

The augmented value of using clinical notes in semi-automated surveillance of deep surgical site infections after colorectal surgery

BACKGROUND

In patients who underwent colorectal surgery, an existing semi-automated surveillance algorithm based on structured data achieves high sensitivity in detecting deep surgical site infections (SSI), however, generates a significant number of false positives. The inclusion of unstructured, clinical narratives to the algorithm may decrease the number of patients requiring manual chart review. The aim of this study was to investigate the performance of this semi-automated surveillance algorithm augmented with a natural language processing (NLP) component to improve positive predictive value (PPV) and thus workload reduction (WR).

METHODS

Retrospective, observational cohort study in patients who underwent colorectal surgery from January 1, 2015, through September 30, 2020. NLP was used to detect keyword counts in clinical notes. Several NLP-algorithms were developed with different count input types and classifiers, and added as component to the original semi-automated algorithm. Traditional manual surveillance was compared with the NLP-augmented surveillance algorithms and sensitivity, specificity, PPV and WR were calculated.

RESULTS

From the NLP-augmented models, the decision tree models with discretized counts or binary counts had the best performance (sensitivity 95.1% (95%CI 83.5-99.4%), WR 60.9%) and improved PPV and WR by only 2.6% and 3.6%, respectively, compared to the original algorithm.

CONCLUSIONS

The addition of an NLP component to the existing algorithm had modest effect on WR (decrease of 1.4-12.5%), at the cost of sensitivity. For future implementation it will be a trade-off between optimal case-finding techniques versus practical considerations such as acceptability and availability of resources.

Predicting sepsis onset using a machine learned causal probabilistic network algorithm based on electronic health records data

Sepsis is a leading cause of mortality and early identification improves survival. With increasing digitalization of health care data automated sepsis prediction models hold promise to aid in prompt recognition. Most previous studies have focused on the intensive care unit (ICU) setting. Yet only a small proportion of sepsis develops in the ICU and there is an apparent clinical benefit to identify patients earlier in the disease trajectory. In this cohort of 82,852 hospital admissions and 8038 sepsis episodes classified according to the Sepsis-3 criteria, we demonstrate that a machine learned score can predict sepsis onset within 48 h using sparse routine electronic health record data outside the ICU. Our score was based on a causal probabilistic network model-SepsisFinder-which has similarities with clinical reasoning. A prediction was generated hourly on all admissions, providing a new variable was registered. Compared to the National Early Warning Score (NEWS2), which is an established method to identify sepsis, the SepsisFinder triggered earlier and had a higher area under receiver operating characteristic curve (AUROC) (0.950 vs. 0.872), as well as area under precision-recall curve (APR) (0.189 vs. 0.149). A machine learning comparator based on a gradient-boosting decision tree model had similar AUROC (0.949) and higher APR (0.239) than SepsisFinder but triggered later than both NEWS2 and SepsisFinder. The precision of SepsisFinder increased if screening was restricted to the earlier admission period and in episodes with bloodstream infection. Furthermore, the SepsisFinder signaled median 5.5 h prior to antibiotic administration. Identifying a high-risk population with this method could be used to tailor clinical interventions and improve patient care.

The accuracy of fully automated algorithms for surveillance of healthcare-associated urinary tract infections in hospitalized patients

BACKGROUND

Surveillance for healthcare-associated infections such as healthcare-associated urinary tract infections (HA-UTI) is important for directing resources and evaluating interventions. However, traditional surveillance methods are resource-intensive and subject to bias.

AIM

To develop and validate a fully automated surveillance algorithm for HA-UTI using electronic health record (EHR) data.

METHODS

Five algorithms were developed using EHR data from 2979 admissions at Karolinska University Hospital from 2010 to 2011: (1) positive urine culture (UCx); (2) positive UCx + UTI codes (International Statistical Classification of Diseases and Related Health Problems, 10^th revision); (3) positive UCx + UTI-specific antibiotics; (4) positive UCx + fever and/or UTI symptoms; (5) algorithm 4 with negation for fever without UTI symptoms. Natural language processing (NLP) was used for processing free-text medical notes. The algorithms were validated in 1258 potential UTI episodes from January to March 2012 and results extrapolated to all UTI episodes within this period (N = 16,712). The reference standard for HA-UTIs was manual record review according to the European Centre for Disease Prevention and Control (and US Centers for Disease Control and Prevention) definitions by trained healthcare personnel.

FINDINGS

Of the 1258 UTI episodes, 163 fulfilled the ECDC HA-UTI definition and the algorithms classified 391, 150, 189, 194, and 153 UTI episodes, respectively, as HA-UTI. Algorithms 1, 2, and 3 had insufficient performances. Algorithm 4 achieved better performance and algorithm 5 performed best for surveillance purposes with sensitivity 0.667 (95% confidence interval: 0.594-0.733), specificity 0.997 (0.996-0.998), positive predictive value 0.719 (0.624-0.807) and negative predictive value 0.997 (0.996-0.997).

CONCLUSION

A fully automated surveillance algorithm based on NLP to find UTI symptoms in free-text had acceptable performance to detect HA-UTI compared to manual record review. Algorithms based on administrative and microbiology data only were not sufficient.

Validation of automated sepsis surveillance based on the Sepsis-3 clinical criteria against physician record review in a general hospital population: observational study using electronic health records data

BACKGROUND

Surveillance of sepsis incidence is important for directing resources and evaluating quality-of-care interventions. The aim was to develop and validate a fully-automated Sepsis-3 based surveillance system in non-intensive care wards using electronic health record (EHR) data, and demonstrate utility by determining the burden of hospital-onset sepsis and variations between wards.

METHODS

A rule-based algorithm was developed using EHR data from a cohort of all adult patients admitted at an academic centre between July 2012 and December 2013. Time in intensive care units was censored. To validate algorithm performance, a stratified random sample of 1000 hospital admissions (674 with and 326 without suspected infection) was classified according to the Sepsis-3 clinical criteria (suspected infection defined as having any culture taken and at least two doses of antimicrobials administered, and an increase in Sequential Organ Failure Assessment (SOFA) score by >2 points) and the likelihood of infection by physician medical record review.

RESULTS

In total 82 653 hospital admissions were included. The Sepsis-3 clinical criteria determined by physician review were met in 343 of 1000 episodes. Among them, 313 (91%) had possible, probable or definite infection. Based on this reference, the algorithm achieved sensitivity 0.887 (95% CI: 0.799 to 0.964), specificity 0.985 (95% CI: 0.978 to 0.991), positive predictive value 0.881 (95% CI: 0.833 to 0.926) and negative predictive value 0.986 (95% CI: 0.973 to 0.996). When applied to the total cohort taking into account the sampling proportions of those with and without suspected infection, the algorithm identified 8599 (10.4%) sepsis episodes. The burden of hospital-onset sepsis (>48 hour after admission) and related in-hospital mortality varied between wards.

CONCLUSIONS

A fully-automated Sepsis-3 based surveillance algorithm using EHR data performed well compared with physician medical record review in non-intensive care wards, and exposed variations in hospital-onset sepsis incidence between wards.

Detecting Protected Health Information in Heterogeneous Clinical Notes

To enable secondary use of healthcare data in a privacy-preserving manner, there is a need for methods capable of automatically identifying protected health information (PHI) in clinical text. To that end, learning predictive models from labeled examples has emerged as a promising alternative to rule-based systems. However, little is known about differences with respect to PHI prevalence in different types of clinical notes and how potential domain differences may affect the performance of predictive models trained on one particular type of note and applied to another. In this study, we analyze the performance of a predictive model trained on an existing PHI corpus of Swedish clinical notes and applied to a variety of clinical notes: written (i) in different clinical specialties, (ii) under different headings, and (iii) by persons in different professions. The results indicate that domain adaption is needed for effective detection of PHI in heterogeneous clinical notes.

Improving Terminology Mapping in Clinical Text with Context-Sensitive Spelling Correction

The mapping of unstructured clinical text to an ontology facilitates meaningful secondary use of health records but is non-trivial due to lexical variation and the abundance of misspellings in hurriedly produced notes. Here, we apply several spelling correction methods to Swedish medical text and evaluate their impact on SNOMED CT mapping; first in a controlled evaluation using medical literature text with induced errors, followed by a partial evaluation on clinical notes. It is shown that the best-performing method is context-sensitive, taking into account trigram frequencies and utilizing a corpus-based dictionary.

Automated Diagnosis Coding with Combined Text Representations

Automated diagnosis coding can be provided efficiently by learning predictive models from historical data; however, discriminating between thousands of codes while allowing a variable number of codes to be assigned is extremely difficult. Here, we explore various text representations and classification models for assigning ICD-9 codes to discharge summaries in MIMIC-III. It is shown that the relative effectiveness of the investigated representations depends on the frequency of the diagnosis code under consideration and that the best performance is obtained by combining models built using different representations.

Prevalence Estimation of Protected Health Information in Swedish Clinical Text

Obscuring protected health information (PHI) in the clinical text of health records facilitates the secondary use of healthcare data in a privacy-preserving manner. Although automatic de-identification of clinical text using machine learning holds much promise, little is known about the relative prevalence of PHI in different types of clinical text and whether there is a need for domain adaptation when learning predictive models from one particular domain and applying it to another. In this study, we address these questions by training a predictive model and using it to estimate the prevalence of PHI in clinical text written (1) in different clinical specialties, (2) in different types of notes (i.e., under different headings), and (3) by persons in different professional roles. It is demonstrated that the overall PHI density is 1.57%; however, substantial differences exist across domains.

Expansion of medical vocabularies using distributional semantics on Japanese patient blogs

Background

Research on medical vocabulary expansion from large corpora has primarily been conducted using text written in English or similar languages, due to a limited availability of large biomedical corpora in most languages. Medical vocabularies are, however, essential also for text mining from corpora written in other languages than English and belonging to a variety of medical genres. The aim of this study was therefore to evaluate medical vocabulary expansion using a corpus very different from those previously used, in terms of grammar and orthographics, as well as in terms of text genre. This was carried out by applying a method based on distributional semantics to the task of extracting medical vocabulary terms from a large corpus of Japanese patient blogs.

Methods

Distributional properties of terms were modelled with random indexing, followed by agglomerative hierarchical clustering of 3 ×100 seed terms from existing vocabularies, belonging to three semantic categories: Medical Finding, Pharmaceutical Drug and Body Part. By automatically extracting unknown terms close to the centroids of the created clusters, candidates for new terms to include in the vocabulary were suggested. The method was evaluated for its ability to retrieve the remaining n terms in existing medical vocabularies.

Results

Removing case particles and using a context window size of 1+1 was a successful strategy for Medical Finding and Pharmaceutical Drug, while retaining case particles and using a window size of 8+8 was better for Body Part. For a 10n long candidate list, the use of different cluster sizes affected the result for Pharmaceutical Drug, while the effect was only marginal for the other two categories. For a list of top n candidates for Body Part, however, clusters with a size of up to two terms were slightly more useful than larger clusters. For Pharmaceutical Drug, the best settings resulted in a recall of 25 % for a candidate list of top n terms and a recall of 68 % for top 10n. For a candidate list of top 10n candidates, the second best results were obtained for Medical Finding: a recall of 58 %, compared to 46 % for Body Part. Only taking the top n candidates into account, however, resulted in a recall of 23 % for Body Part, compared to 16 % for Medical Finding.

Conclusions

Different settings for corpus pre-processing, window sizes and cluster sizes were suitable for different semantic categories and for different lengths of candidate lists, showing the need to adapt parameters, not only to the language and text genre used, but also to the semantic category for which the vocabulary is to be expanded. The results show, however, that the investigated choices for pre-processing and parameter settings were successful, and that a Japanese blog corpus, which in many ways differs from those used in previous studies, can be a useful resource for medical vocabulary expansion.

Learning temporal weights of clinical events using variable importance

Background

Longitudinal data sources, such as electronic health records (EHRs), are very valuable for monitoring adverse drug events (ADEs). However, ADEs are heavily under-reported in EHRs. Using machine learning algorithms to automatically detect patients that should have had ADEs reported in their health records is an efficient and effective solution. One of the challenges to that end is how to take into account the temporality of clinical events, which are time stamped in EHRs, and providing these as features for machine learning algorithms to exploit. Previous research on this topic suggests that representing EHR data as a bag of temporally weighted clinical events is promising; however, the weights were in that case pre-assigned according to their time stamps, which is limited and potentially less accurate. This study therefore focuses on how to learn weights that effectively take into account the temporality and importance of clinical events for ADE detection.

Methods

Variable importance obtained from the random forest learning algorithm is used for extracting temporal weights. Two strategies are proposed for applying the learned weights: weighted aggregation and weighted sampling. The first strategy aggregates the weighted clinical events from different time windows to form new features; the second strategy retains the original features but samples them by using their weights as probabilities when building each tree in the forest. The predictive performance of random forest models using the learned weights with the two strategies is compared to using pre-assigned weights. In addition, to assess the sensitivity of the weight-learning procedure, weights from different granularity levels are evaluated and compared.

Results

In the weighted sampling strategy, using learned weights significantly improves the predictive performance, in comparison to using pre-assigned weights; however, there is no significant difference between them in the weighted aggregation strategy. Moreover, the granularity of the weight learning procedure has a significant impact on the former, but not on the latter.

Conclusions

Learning temporal weights is significantly beneficial in terms of predictive performance with the weighted sampling strategy. Moreover, weighted aggregation generally diminishes the impact of temporal weighting of the clinical events, irrespective of whether the weights are pre-assigned or learned.

Ensembles of randomized trees using diverse distributed representations of clinical events

BACKGROUND

Learning deep representations of clinical events based on their distributions in electronic health records has been shown to allow for subsequent training of higher-performing predictive models compared to the use of shallow, count-based representations. The predictive performance may be further improved by utilizing multiple representations of the same events, which can be obtained by, for instance, manipulating the representation learning procedure. The question, however, remains how to make best use of a set of diverse representations of clinical events - modeled in an ensemble of semantic spaces - for the purpose of predictive modeling.

METHODS

Three different ways of exploiting a set of (ten) distributed representations of four types of clinical events - diagnosis codes, drug codes, measurements, and words in clinical notes - are investigated in a series of experiments using ensembles of randomized trees. Here, the semantic space ensembles are obtained by varying the context window size in the representation learning procedure. The proposed method trains a forest wherein each tree is built from a bootstrap replicate of the training set whose entire original feature set is represented in a randomly selected set of semantic spaces - corresponding to the considered data types - of a given context window size.

RESULTS

The proposed method significantly outperforms concatenating the multiple representations of the bagged dataset; it also significantly outperforms representing, for each decision tree, only a subset of the features in a randomly selected set of semantic spaces. A follow-up analysis indicates that the proposed method exhibits less diversity while significantly improving average tree performance. It is also shown that the size of the semantic space ensemble has a significant impact on predictive performance and that performance tends to improve as the size increases.

CONCLUSIONS

The strategy for utilizing a set of diverse distributed representations of clinical events when constructing ensembles of randomized trees has a significant impact on predictive performance. The most successful strategy - significantly outperforming the considered alternatives - involves randomly sampling distributed representations of the clinical events when building each decision tree in the forest.

Predictive modeling of structured electronic health records for adverse drug event detection

Background

The digitization of healthcare data, resulting from the increasingly widespread adoption of electronic health records, has greatly facilitated its analysis by computational methods and thereby enabled large-scale secondary use thereof. This can be exploited to support public health activities such as pharmacovigilance, wherein the safety of drugs is monitored to inform regulatory decisions about sustained use. To that end, electronic health records have emerged as a potentially valuable data source, providing access to longitudinal observations of patient treatment and drug use. A nascent line of research concerns predictive modeling of healthcare data for the automatic detection of adverse drug events, which presents its own set of challenges: it is not yet clear how to represent the heterogeneous data types in a manner conducive to learning high-performing machine learning models.

Methods

Datasets from an electronic health record database are used for learning predictive models with the purpose of detecting adverse drug events. The use and representation of two data types, as well as their combination, are studied: clinical codes, describing prescribed drugs and assigned diagnoses, and measurements. Feature selection is conducted on the various types of data to reduce dimensionality and sparsity, while allowing for an in-depth feature analysis of the usefulness of each data type and representation.

Results

Within each data type, combining multiple representations yields better predictive performance compared to using any single representation. The use of clinical codes for adverse drug event detection significantly outperforms the use of measurements; however, there is no significant difference over datasets between using only clinical codes and their combination with measurements. For certain adverse drug events, the combination does, however, outperform using only clinical codes. Feature selection leads to increased predictive performance for both data types, in isolation and combined.

Conclusions

We have demonstrated how machine learning can be applied to electronic health records for the purpose of detecting adverse drug events and proposed solutions to some of the challenges this presents, including how to represent the various data types. Overall, clinical codes are more useful than measurements and, in specific cases, it is beneficial to combine the two.

Handling Temporality of Clinical Events for Drug Safety Surveillance

Using longitudinal data in electronic health records (EHRs) for post-marketing adverse drug event (ADE) detection allows for monitoring patients throughout their medical history. Machine learning methods have been shown to be efficient and effective in screening health records and detecting ADEs. How best to exploit historical data, as encoded by clinical events in EHRs is, however, not very well understood. In this study, three strategies for handling temporality of clinical events are proposed and evaluated using an EHR database from Stockholm, Sweden. The random forest learning algorithm is applied to predict fourteen ADEs using clinical events collected from different lengths of patient history. The results show that, in general, including longer patient history leads to improved predictive performance, and that assigning weights to events according to time distance from the ADE yields the biggest improvement.

Identifying adverse drug event information in clinical notes with distributional semantic representations of context

For the purpose of post-marketing drug safety surveillance, which has traditionally relied on the voluntary reporting of individual cases of adverse drug events (ADEs), other sources of information are now being explored, including electronic health records (EHRs), which give us access to enormous amounts of longitudinal observations of the treatment of patients and their drug use. Adverse drug events, which can be encoded in EHRs with certain diagnosis codes, are, however, heavily underreported. It is therefore important to develop capabilities to process, by means of computational methods, the more unstructured EHR data in the form of clinical notes, where clinicians may describe and reason around suspected ADEs. In this study, we report on the creation of an annotated corpus of Swedish health records for the purpose of learning to identify information pertaining to ADEs present in clinical notes. To this end, three key tasks are tackled: recognizing relevant named entities (disorders, symptoms, drugs), labeling attributes of the recognized entities (negation, speculation, temporality), and relationships between them (indication, adverse drug event). For each of the three tasks, leveraging models of distributional semantics - i.e., unsupervised methods that exploit co-occurrence information to model, typically in vector space, the meaning of words - and, in particular, combinations of such models, is shown to improve the predictive performance. The ability to make use of such unsupervised methods is critical when faced with large amounts of sparse and high-dimensional data, especially in domains where annotated resources are scarce.

Synonym extraction and abbreviation expansion with ensembles of semantic spaces

Background

Terminologies that account for variation in language use by linking synonyms and abbreviations to their corresponding concept are important enablers of high-quality information extraction from medical texts. Due to the use of specialized sub-languages in the medical domain, manual construction of semantic resources that accurately reflect language use is both costly and challenging, often resulting in low coverage. Although models of distributional semantics applied to large corpora provide a potential means of supporting development of such resources, their ability to isolate synonymy from other semantic relations is limited. Their application in the clinical domain has also only recently begun to be explored. Combining distributional models and applying them to different types of corpora may lead to enhanced performance on the tasks of automatically extracting synonyms and abbreviation-expansion pairs.

Results

A combination of two distributional models – Random Indexing and Random Permutation – employed in conjunction with a single corpus outperforms using either of the models in isolation. Furthermore, combining semantic spaces induced from different types of corpora – a corpus of clinical text and a corpus of medical journal articles – further improves results, outperforming a combination of semantic spaces induced from a single source, as well as a single semantic space induced from the conjoint corpus. A combination strategy that simply sums the cosine similarity scores of candidate terms is generally the most profitable out of the ones explored. Finally, applying simple post-processing filtering rules yields substantial performance gains on the tasks of extracting abbreviation-expansion pairs, but not synonyms. The best results, measured as recall in a list of ten candidate terms, for the three tasks are: 0.39 for abbreviations to long forms, 0.33 for long forms to abbreviations, and 0.47 for synonyms.

Conclusions

This study demonstrates that ensembles of semantic spaces can yield improved performance on the tasks of automatically extracting synonyms and abbreviation-expansion pairs. This notion, which merits further exploration, allows different distributional models – with different model parameters – and different types of corpora to be combined, potentially allowing enhanced performance to be obtained on a wide range of natural language processing tasks.

Identifying synonymy between SNOMED clinical terms of varying length using distributional analysis of electronic health records

Medical terminologies and ontologies are important tools for natural language processing of health record narratives. To account for the variability of language use, synonyms need to be stored in a semantic resource as textual instantiations of a concept. Developing such resources manually is, however, prohibitively expensive and likely to result in low coverage. To facilitate and expedite the process of lexical resource development, distributional analysis of large corpora provides a powerful data-driven means of (semi-)automatically identifying semantic relations, including synonymy, between terms. In this paper, we demonstrate how distributional analysis of a large corpus of electronic health records - the MIMIC-II database - can be employed to extract synonyms of SNOMED CT preferred terms. A distinctive feature of our method is its ability to identify synonymous relations between terms of varying length.

A comparative study of the preventative effects exerted by three probiotics, Bifidobacterium lactis, Lactobacillus casei and Lactobacillus acidophilus, in the TNBS model of rat colitis

AIMS

The intestinal anti-inflammatory effects of three probiotics with immunomodulatory properties, Lactobacillus casei, Lactobacillus acidophilus and Bifidobacterium lactis, were evaluated and compared in the trinitrobenzenesulphonic acid (TNBS) model of rat colitis.

METHODS AND RESULTS

Colitis was induced in rats by intracolonic administration of 10 mg of TNBS dissolved in 0.25 ml of 50% ethanol. Each probiotic was administered orally (5x10(8) CFU suspended in 0.5 ml of skimmed milk) for 3 weeks, starting 2 weeks before the administration of TNBS. Colonic damage was evaluated histologically and biochemically 1 week after TNBS instillation. The results obtained revealed that all probiotics assayed showed intestinal anti-inflammatory effects, macroscopically evidenced by a significant reduction in the colonic weight/length ratio. Only B. lactis showed a lower incidence of diarrhoea in comparison with untreated rats. Biochemically, all probiotics restored colonic glutathione levels, depleted as a consequence of the oxidative stress of the inflammatory process. Bifidobacterium lactis treatment reduced colonic tumour necrosis factor (TNF)-alpha production, and inducible nitric oxide synthase (iNOS) and cyclo-oxygenase-2 (COX-2) expression; L. acidophilus administration reduced colonic leukotriene B4 production and iNOS expression and L. casei intake was associated with a decrease in colonic COX-2 expression.

CONCLUSION

The three probiotics assayed have shown intestinal anti-inflammatory activity in the TNBS model of rat colitis, although each probiotic shows its own anti-inflammatory profile.

SIGNIFICANCE AND IMPACT OF THE STUDY

These probiotics could be considered as potential adjuvants in the treatment of inflammatory bowel disease, although more studies are required in order to demonstrate their efficacy in humans.

Prebiotics enhance survival and prolong the retention period of specific probiotic inocula in an in vivo murine model

AIM

To identify novel prebiotics that could be used to maintain persistence of three representative probiotic strains in vivo.

METHODS AND RESULTS

Test mice were treated with prebiotics soybean oligosaccharide (SOS), fructooligosaccharide (FOS) or inulin, followed by probiotics Lactobacillus acidophilus LAFTI L10 (L10), Bifidobacterium lactis LAFTI B94 (B94) or Lactobacillus casei L26 LAFTI (L26). Faecal samples were then collected and analysed using selective medium and PCR analysis to determine the presence of the probiotic strains. In contrast to the control groups, in mice fed prebiotics, the survival and retention time of the test probiotics was increased extensively. SOS and FOS prolonged the retention period of L10 from 24 to 30 h. Of the three prebiotics, FOS gave the best result with B94, prolonging the retention period from 3 to > or =10 days. Of the three prebiotics, inulin gave the best result for L26, prolonging the retention period from 2 to > or =6 days.

CONCLUSIONS

The prebiotics SOS, FOS and inulin significantly enhance survival and prolong the retention period of L10, B94 and L26 in vivo.

SIGNIFICANCE AND IMPACT OF THE STUDY

Our results demonstrate the potential use of FOS, inulin and SOS as prebiotics in conjunction with the probiotic strains L10, B94 and L26 for new synbiotic products.

Rheological properties and sensory characteristics of set-type soy yogurt

The study examined chemical composition and rheological and sensory properties of probiotic soy yogurt during 28 day storage at 4 degrees C. Soymilk supplemented with 2% (w/v) inulin or 1% (w/v) each of raffinose and glucose was used as a base for soy yogurt manufacture. Viability of probiotic organisms and their metabolic activity measured as production of organic acids and aldehyde content responsible for beany flavor, as well as rheological and sensory properties of soy yogurt, were examined. Inulin or raffinose/glucose supplementation in soymilk increased the bacterial population by one log cycle and the amount of lactic acid. Probiotic bacteria metabolized more aldehyde than yogurt culture and substantially reduced the beaniness in soy yogurt as determined by sensory evaluation. The probiotic soy yogurts showed more viscous and pseudoplastic properties than the control soy yogurts, but the sensory evaluation results showed preference for the control soy yogurts which were slightly less viscous. Control soy yogurt provided better mouth feel than probiotic soy yogurts.

Survival and retention of the probiotic Lactobacillus casei LAFTI�L26 in the gastrointestinal tract of the mouse

AIMS

This study aimed to develop methods for the detection of the probiotic Lactobacillus casei LAFTI L26 (L26) from mouse faeces, and to determine the survival and retention time of L26 in the mouse gastrointestinal tract.

METHODS AND RESULTS

A selective medium, de Man Rogosa Sharpe (MRS) + bromocresol green + vancomycin (MGV), was designed for the isolation and enumeration of L26 from faecal samples of mice. PCR primers were designed to confirm the identity of L26-like colonies on MGV. These primers did not produce PCR products from related organisms that grew on MGV. Following the administration of L26 to BALB/c mice, faecal samples were collected and analysed using the designed methods. Survival studies showed viable L26 cells to be present in the faeces of mice for >48 h.

CONCLUSIONS

Our results suggest that L26 is able to survive and be retained within the digestive tract of mice for at least 48 h following oral administration.

SIGNIFICANCE AND IMPACT OF THE STUDY

MGV allows effective recovery of L26 from the background microbiota, including lactobacilli of mice. PCR was used to confirm that L26-like colonies were correctly identified as L26. Given the long retention time of L26 in the gastrointestinal tract of mice, it would appear that this probiotic strain may survive in the human gastrointestinal tract.

Reversal in fatigued athletes of a defect in interferon gamma secretion after administration of Lactobacillus acidophilus

BACKGROUND

Fatigue and impaired performance in athletes is well recognised and has been loosely linked to "overtraining". Reduced concentration of IgA in the saliva and increased shedding of Epstein Barr virus (EBV) have been associated with intense training in elite athletes.

OBJECTIVE

To determine whether athletes presenting with fatigue and impaired performance had an immune defect relevant to defective containment of EBV infection, and whether a probiotic preparation (Lactobacillus acidophilus) shown to enhance mucosal immunity in animal models could reverse any detected abnormality.

RESULTS

The fatigued athletes had clinical characteristics consistent with re-activation of EBV infection and significantly (p = 0.02) less secretion of interferon (IFN) gamma from blood CD4 positive T cells. After one month of daily capsules containing 2 x 10(10) colony forming units of L acidophilus, secretion of IFNgamma from T cells had increased significantly (p = 0.01) to levels found in healthy control athletes. A significant (p = 0.03) increase in salivary IFNgamma concentrations in healthy control athletes after the one month course of L acidophilus demonstrated in man the capacity for this probiotic to enhance the mucosal IFNgamma concentration.

CONCLUSION

This is the first evidence of a T cell defect in fatigued athletes, and of its reversal following probiotic therapy.

Probiotic research in Australia, New Zealand and the Asia-Pacific region

Although the epicentres of probiotic research in the past decade have been Japan and Europe, researchers in the Asia-Pacific region have actively contributed to the growing understanding of the intestinal microbial ecosystem, and interactions between gut bacteria, diet and health of the human host. A number of new probiotic strains have been developed in the region that have been demonstrated to have beneficial impacts on health in animal and human trials, including improved protection against intestinal pathogens and modulation of the immune system. Probiotics targeted to animals, including aquaculture, feature heavily in many Asian countries. Developments in probiotic technologies have included microencapsulation techniques, antimicrobial production in fermented meats, and synbiotic combinations. In particular, the impact of resistant starch on the intestinal environment and fermentation by intestinal bacteria has been intensively studied and new probiotic strains selected specifically for synbiotic combinations with resistant starch. This paper provides an overview of probiotic research within Australia, New Zealand and a number of Asian countries, and lists scientists in the Asia-Pacific region involved in various aspects of probiotic research and development.

The effect of processed meat and meat starter cultures on gastrointestinal colonization and virulence of Listeria monocytogenes in mice

Listeria monocytogenes is a foodborne pathogen of major concern to the food industry in general and the meat industry in particular. The aim of this study was firstly to identify a strain of Listeria that was virulent in SPF BALB/c mice. Secondly, to investigate if a traditional meat starter culture (FloraCarn) and nontraditional meat starter (NTMS) cultures of dairy product and human origin (Lactobacillus and bifidobacteria) inhibit this pathogen in vivo. In addition, the inhibition of Listeria was investigated in vitro. In vitro inhibition was investigated using an agar inhibition assay, where soft agar containing the pathogen was laid over colonies of NTMS cultures, and inhibition expressed as the zones of inhibition developing around the colonies. For assessment of virulence, mice were intragastrically challenged with broth cultures of five strains of Listeria. For assessment of anti-listeria effect in vivo, the Listeria strain proven to be most pathogenic (LM3) was given to mice in salami batter containing no other added cultures (control) or batter inoculated with either (1) FloraCarn, (2) a NTMS culture, or (3) a combination of FloraCarn and a NTMS culture. The batter was given to mice after a 3-day fermentation and faecal levels of pathogen and body weight were monitored. Intragastric challenge with LM3, but no other strains, resulted in a significant weight loss (p<0.05) and up to 10(6) colony forming units (cfu) of LM3 per gram faeces. No weight loss was observed in animals fed with salami batter containing LM3. Consumption of salami batter fermented by a combination of NTMS culture (Lactobacillus acidophilus LAFTI(R) L10) and FloraCarn reduced faecal levels of the pathogen by 2.5 log units compared to the control. Consumption of salami batter fermented with FloraCarn and LAFTI(R) L10 (L10) alone reduced faecal levels by 0.5-1 and 1.5 log units, respectively. Of the NTMS cultures investigated here, L10 displayed the greatest inhibition of LM3 in vitro. These results indicate that the ability of pathogenic Listeria to cause listeriosis is dependent on the nature of the food in which the pathogen is present, and that a traditional meat starter culture (FloraCarn) and some NTMS cultures, particularly L10, inhibit growth of the pathogen during passage through the gastrointestinal tract.

Development of the LAFTI range of probiotic cultures

Probiotics are mono or mixed cultures of live microorganisms which, when applied to man or animal, beneficially affect the host by improving the properties of the indigenous microflora. Some of the beneficial effects that a probiotic culture can have on its host include improved digestion and absorption of various nutrients (e.g. lactose, starch), production of vitamins and growth factors, protection against pathogens, stimulation of the immune response, reduction of cholesterol levels and reduction of diarrhoea.

Application of nontraditional meat starter cultures in production of Hungarian salami

Listeria monocytogenes and Escherichia coli O111 have been implicated in several outbreaks of food-borne disease linked to smallgoods products. Traditional meat starter cultures, containing a mixture of lactic acid bacteria (LAB) and staphylococci, are used to maintain safety and sensory properties of Hungarian salami. The present study investigated if nontraditional meat starter (NTMS) cultures can be used for improving the safety of Hungarian salami. Salami batter was inoculated with List. monocytogenes and E. coli and subsequently fermented with NTMS cultures and a commercially available meat starter. A total of 15 NTMS cultures were tested. The salami was monitored for levels of pathogen, LAB and pH. When used in conjunction with the commercial meat starter, 9 NTMS cultures reduced the E. coli O111 count by more than 2.5 log units, whereas 10 of the NTMS cultures reduced List. monocytogenes by more than 2.5 log units. The commercial meat starter alone reduced E. coli and List. monocytogenes by 1.2 and 1.3 log units, respectively. Some NTMS cultures reduced the pathogen count without affecting pH of the salami batter. All NTMS cultures survived in salami throughout fermentation and maturation. It was concluded that NTMS cultures, including Lactobacillus acidophilus LAFTI L10, L. paracasei LAFTI L26, L. paracasei 5119, Lactobacillus sp. L24 and Bifidobacterium lactis LAFTI B94, may be used to increase the safety of Hungarian salami because these cultures gave strong inhibition of both E. coli O111 and List. monocytogenes.

Isolation of human faecal bifidobacteria which reduce signs of Salmonella infection when orogastrically dosed to mice

AIMS

The aim of the study was to isolate human bifidobacteria that inhibit growth of Salmonella typhimurium in vitro, and provide protection against Salmonella infection in mice.

METHODS AND RESULTS

A total of 92 micro-organisms, which displayed antagonist activity against Salm. typhimurium in vitro, were isolated from human faecal material. Based on their Gram stain status, cultures were pooled and tested for anti-Salmonella activity. The Gram-variable group was the most active. From that group, three bifidobacteria (Laftitrade markB22, B74 and B97) individually showed good pathogen inhibition in vivo.

CONCLUSION

Oral administration of certain human bifidobacteria provides protection against Salmonella infection in mice.

SIGNIFICANCE AND IMPACT OF THE STUDY

These results indicate that certain bifidobacteria may be used as a prophylaxis for reduced incidence and severity of Salmonella infections.

Adrenal steroid dysregulation in dystrophia myotonica

OBJECTIVE

To evaluate circulating adrenal steroid hormones, cortisol diurnal rhythm and the negative feedback function of the cortisol axis in patients with dystrophia myotonica (DyM), a disease where metabolic disturbances, peripheral insulin insensitivity and cognitive dysfunction are common features.

DESIGN

Morning serum levels of dehydroepiandrosterone sulphate, androstenedione, 17 alpha-hydroxy progesterone and cortisol; morning serum levels of testosterone and insulin; diurnal rhythm of saliva cortisol; and an overnight dexamethasone suppression test, together with a cognitive screening test in men with DyM and in controls.

SETTING

Outpatient clinic in co-operation with Umeå University Hospital.

SUBJECTS

Fifteen men with DyM and 13 age-matched controls.

MAIN OUTCOME MEASURES

Adrenal steroid hormone levels, diurnal rhythm of saliva cortisol, dexamethasone suppression test and Mini Mental State Examination scores.

RESULTS

Morning serum levels of dehydroepiandrosterone sulphate, androstenedione and 17 alpha-hydroxy progesterone were significantly decreased in DyM after inclusion of age and body mass index in multiple regression analyses (48, 26 and 32% decreases, respectively). An abnormal diurnal rhythm of saliva cortisol was present in all patients, mean saliva cortisol levels being significantly increased (33%) in DyM patients. Dexamethasone suppressibility did not differ between groups. DyM patients scored significantly lower on the Mini Mental State Examination (P < 0.001).

CONCLUSIONS

These results indicate an abnormal adrenal steroid hormone secretion in DyM, which may contribute to peripheral insulin sensitivity as well as cognitive impairment in these patients.

Ciprofloxacin induces an immunomodulatory stress response in human T lymphocytes

Exposure of cells to adverse environmental conditions invokes a genetically programmed series of events resulting in the induction of specific genes. The fluoroquinolone antibiotic ciprofloxacin has recently been reported to upregulate interleukin-2 (IL-2) gene induction. In the present investigation, the effect of ciprofloxacin at supratherapeutic concentrations on immediate-early (<2 h) gene expression in primary human peripheral blood lymphocytes was studied with Northern blots. In addition, transcriptional activity of IL-2 and metallothionein enhancer and promoter regions and transcription factors AP-1, NF-kappaB, and NF-AT were analyzed by chloramphenicol acetyltransferase (CAT) and electrophoretic mobility shift assays, respectively. The concentration of c-fos, c-jun, c-myc, junB, and fra-1 mRNAs was increased in activated peripheral blood lymphocytes incubated with ciprofloxacin compared to that in untreated controls. Ciprofloxacin increased CAT activity in stimulated lymphocytes transfected with plasmids containing either the IL-2 or metallothionein enhancer. Furthermore, among the transcription factors tested, AP-1 activity was increased in stimulated purified T helper lymphocytes incubated with ciprofloxacin compared to drug-free controls. Taken together, ciprofloxacin increased the levels of immediate-early transcripts, enhanced IL-2 and metallothionein promoter induction, and upregulated AP-1 concentrations in primary lymphocytes, reflecting a program commonly observed in mammalian stress responses.

Abnormal DNA damage-inducible protein in cells from Sjögren's syndrome patients

Antinuclear antibodies are commonly found in patients with Sjögren's syndrome. It has been suggested that the development of antinuclear antibodies depends on the activation of the spliceosome and other transcription-related subcellular particles, some of which have recently been shown also to function in DNA-modifying processes, such as DNA repair and V(D)J recombination. These observations add weight to a previously proposed model for the aetiology of Sjögren's syndrome. This includes the abnormal processing of the T-cell receptor and immunoglobulin genes. To test this hypothesis further, the present study on DNA-modifying proteins in Sjögren's syndrome was initiated. Gel-shift experiments using protein extracted from UV-treated Sjögren cells provided evidence of high molecular weight DNA-binding protein in six out of 12 Sjögren patients studied (but not among seven healthy controls). Some Sjögren sera displayed antibodies to protein extracts from cells treated with psoralen plus UVA radiation. These results indicate an abnormal DNA damage-inducible response in Sjögren's syndrome. It may therefore be concluded that alterations in nuclear protein may play a role in the aetiology of Sjögren's syndrome.

How does sumatriptan nasal spray perform in clinical practice?

Migraineurs (94F, 24M) with previous experience of subcutaneous sumatriptan and/or sumatriptan tablets were asked their opinion on sumatriptan nasal spray treatment particularly with respect to onset of action, total efficacy, tolerability, and user friendliness. The information was obtained by means of a self-administered questionnaire handed out at the time of prescription of the nasal spray. The results are based on the patients' cumulative experience of having treated at least two migraine attacks with the spray (20 mg). Sumatriptan nasal spray (20 mg) was perceived to have a faster onset of action and, with the exception of a bad taste, to have a better tolerability than the tablets. Compared with subcutaneous sumatriptan, the nasal spray was claimed to be less effective in reducing symptoms of migraine attacks but to cause fewer adverse events. A bitter taste was the most commonly reported side effect of sumatriptan nasal spray--reported by 68% (80 out of 118) of the migraineurs. The user friendliness of sumatriptan nasal spray was rated better than that of subcutaneous sumatriptan and/or sumatriptan tablets. The overall impression of sumatriptan nasal spray was reported to be better or equal to that of the tablet and the injection by 57% and 46%, respectively. It is concluded that the results obtained in clinical practice are very much in line with those obtained in controlled clinical trials. The overall impression of sumatriptan nasal spray is that it is user friendly and useful in the acute treatment of migraine attacks of moderate intensity.

Identification and phylogenetic analysis of Lactobacillus using multiplex RAPD-PCR

Multiplex RAPD-PCR was used to generate unique and identifying DNA profiles for isolates of the genus Lactobacillus. The method that was used was based on the combination of two 10-mer oligonucleotides in a single PCR. The generated RAPD profiles enabled discrimination of all lactobacillus strains that were used in this study. A dendrogram was generated from the RAPD profiles. The results of genetic relatedness obtained from the dendrogram were compared with the results obtained using carbohydrate fermentation profiles. Most of the gastrointestinal isolates studied could not be grouped using carbohydrate fermentation profiles. The RAPD profiles provided sufficient information to prepare a dendrogram of genetic relatedness. The gastrointestinal isolates were clustered together on the dendrogram. Furthermore an isolate originating from the stomach (strain ML004) was closely related to Lactobacillus fermentum. It was concluded that multiplex RAPD-PCR was useful for characterisation and inference of relatedness of Lactobacillus isolates.

Role of the Ada adaptor complex in gene activation by the glucocorticoid receptor

We have shown that the Ada adaptor complex is important for the gene activation capacity of the glucocorticoid receptor in yeast. The recently isolated human Ada2 protein also increases the potency of the receptor protein in mammalian cells. The Ada pathway is of key significance for the tau1 core transactivation domain (tau1c) of the receptor, which requires Ada for activity in vivo and in vitro. Ada2 can be precipitated from nuclear extracts by a glutathione S-transferase-tau1 fusion protein coupled to agarose beads, and a direct interaction between Ada2 and tau1c can be shown by using purified proteins. This interaction is strongly reduced by a mutation in tau1c that reduces transactivation activity. Mutations affecting the Ada complex do not reverse transcriptional squelching by the tau1 domain, as they do for the VP16 transactivation domain, and thus these powerful acidic activators differ in at least some important aspects of gene activation. Mutations that reduce the activity of the tau1c domain in wild-type yeast strains cause similar reductions in ada mutants that contain little or no Ada activity. Thus, gene activation mechanisms, in addition to the Ada pathway, are involved in the activity of the tau1c domain.

Adhesion of Lactobacillus fermentum 104-S to porcine stomach mucus

The adhesion to whole and fractionated porcine gastric mucus of both Lactobacillus fermentum 104-S cells and a saccharide extracted from this strain was investigated. It has been shown previously that this saccharide had affinity for nonsecreting gastric epithelium. The mucus component(s) with affinity the bacterial cells was partly characterized by gel filtration and treatment with protease or metaperiodate. L. fermentum 104-S extracts containing the saccharide were radioactively labeled, fractionated by gel filtration, and tested for affinity for the gastric mucus component showing receptor activity for the whole cells of strain 104-S. The mucus material with affinity for the bacterial cells had a relative molecular weight of 30-70 K. From the results of treatment with protease or metaperiodate, it is proposed that the mucus components(s) that adhered to the whole bacterial cells contained glycoprotein groups. The radioactively labeled saccharide extracted from L. fermentum 104-S cells did not bind to the mucus fraction that had affinity for the whole cells. Conclusively, we suggest that the mechanism by which cells of L. fermentum 104-S adhere to the gastric mucus is different from the mechanism mediating the adhesion of this strain to the nonsecreting gastric epithelium. Cells of L. fermentum 104-S adhere to a glycoproteinaceous mucus component with a relative molecular weight of 30-70 K.

Probiotic treatment of small intestinal bacterial overgrowth by Lactobacillus fermentum KLD

The principle of using harmless bacteria for conquering pathogens has been used for many years. It has been used prophylactically against travellers' diarrhoea and for protection of recurrent pseudomembranous colitis. The aim of this study was to treat a chronic infectious condition, small intestinal bacterial overgrowth, by oral administration of a certain strain of Lactobacillus. 17 patients with long-standing bacterial overgrowth of the small intestine were included. The study was designed as a double-blind cross-over, where the patients were their own controls. The study was divided into 4 parts. (A) For the first 2 weeks placebo was given b.i.d. (B) For the next 4 weeks patients received either placebo or 10(10) Lactobacillus fermentum KLD b.i.d. (C) A wash-out period of 4 weeks followed. (D) Finally, for the second 4 week treatment period patients were crossed over to receive either lactobacilli or placebo. A hydrogen breath test with 50 g glucose was performed at the start and at the end of each period. Symptom scores were recorded on the last week of each period. The study was completed by 14 patients. Lactobacillus treatment showed no significant difference compared to placebo with respect to the results of the hydrogen breath test: 29 (3-95) vs 14 (3-129) ppm, (median and 10th and 90th percentiles), stool frequency: 14 (8-40) vs 12 (7-31) defecations/week. or symptom score: 12 (5-46) vs 17 (6-42) scores/week). High numbers of L. fermentum KLD in faecal samples were only seen in 2 patients. In conclusion, dosage with L. fermentum KLD in this study did not significantly alter the parameters investigated.