Transformation from zero tolerance to living with COVID-19 in New Taipei City, Taiwan. Experience of the FEMH "home-hotel-hospital" care model

In March 2022, local cases of COVID-19 infections of the Omicron variant were identified in Taiwan. In response to impending community transmission, the "Home-Hotel-Hospital" (3H) care model was implemented by the Far Eastern Memorial Hospital (FEMH). It established the first remote home care center in Taiwan and two quarantine centers in two hotels. The hospital focused on care for critical COVID-19 patients, community screening, and telehealth care. The home care call center evaluated and triaged up to 104,244 cases and provided remote home care for 96,894 cases within the first three months; in 2022, it provided home care to 107,095 patients. The two quarantine hotels admitted a total of 1834 individuals. A total of 3796 COVID-19 patients were admitted to the hospital-367 in intensive care. The telehealth outpatient clinic-including the online video clinic-served 25,775 cases; 21.5% (n = 5544) of them were prescribed oral anti-viral medications. In 2022, the FEMH prescribed oral anti-viral therapies to a total of 12,571 cases. The FEMH 3H care model not only enabled non-critical patients to recover at home, but also provided severely ill patients access to timely in-hospital care. In the future, this model will continue to play a significant role in COVID-19 management.

Emotional disturbance and risk factors among COVID-19 confirmed cases in isolation hotels

Patients with coronavirus disease 2019 (COVID-19) has been isolated in hospital-managed isolation hotels under a policy of the Taiwan government. Centrally isolation patients are more likely to experience psychological symptoms. The purpose of the study was to investigate emotional disturbance during their isolation period and then pinpoint the factors during their isolation period associated with the emotional disturbance. We retrospectively analysed the medical charts of the patients confined to a Banqiao isolation hotel between May 28 and July 3, 2021. The 5-item brief symptom rating scale (BSRS-5) was used to evaluate emotional disturbance levels. Descriptive and logistic regression was used for the data analysis. In total, 197 complete medical records were reviewed, and of these 84 (42.6%) showed emotional disturbance. The majority of them reported only minor disturbance (n = 49, 58.3%). After controlling for confounding factors, being satisfied about medical information was the only protective factor associated with emotional disturbance (OR = 0.2, P = 0.018). Being a male patient (OR = 3.0, P = 0.005), worrying about stigmatization (OR = 2.2, P = 0.041) and being unable to contact family members (OR = 2.9, P = 0.018) increased the risk of experiencing emotional disturbance. Patients with clinical symptoms, namely sore throat (OR = 3.4, P = 0.013) and muscle aches (OR = 6.3, P = 0.005), were also found to be more likely to report emotional disturbance. Mental disturbance commonly occurs among patient with COVID-19 who are isolated in a hospital-managed hotel. Being a male patient, having symptoms, namely a sore throat and muscle pain, being unable to contact family and/or a failure to receive sufficient medical information were found to be associated with emotional disturbance. In order to help isolated patients, government officials should provide a clear rationale for isolation and recognize the patients' efforts to follow the government's policy, which will help to minimize social stigma.

Training a Deep Contextualized Language Model for International Classification of Diseases, 10th Revision Classification via Federated Learning: Model Development and Validation Study

Background

The automatic coding of clinical text documents by using the International Classification of Diseases, 10th Revision (ICD-10) can be performed for statistical analyses and reimbursements. With the development of natural language processing models, new transformer architectures with attention mechanisms have outperformed previous models. Although multicenter training may increase a model’s performance and external validity, the privacy of clinical documents should be protected. We used federated learning to train a model with multicenter data, without sharing data per se.

Objective

This study aims to train a classification model via federated learning for ICD-10 multilabel classification.

Methods

Text data from discharge notes in electronic medical records were collected from the following three medical centers: Far Eastern Memorial Hospital, National Taiwan University Hospital, and Taipei Veterans General Hospital. After comparing the performance of different variants of bidirectional encoder representations from transformers (BERT), PubMedBERT was chosen for the word embeddings. With regard to preprocessing, the nonalphanumeric characters were retained because the model’s performance decreased after the removal of these characters. To explain the outputs of our model, we added a label attention mechanism to the model architecture. The model was trained with data from each of the three hospitals separately and via federated learning. The models trained via federated learning and the models trained with local data were compared on a testing set that was composed of data from the three hospitals. The micro F1 score was used to evaluate model performance across all 3 centers.

Results

The F1 scores of PubMedBERT, RoBERTa (Robustly Optimized BERT Pretraining Approach), ClinicalBERT, and BioBERT (BERT for Biomedical Text Mining) were 0.735, 0.692, 0.711, and 0.721, respectively. The F1 score of the model that retained nonalphanumeric characters was 0.8120, whereas the F1 score after removing these characters was 0.7875—a decrease of 0.0245 (3.11%). The F1 scores on the testing set were 0.6142, 0.4472, 0.5353, and 0.2522 for the federated learning, Far Eastern Memorial Hospital, National Taiwan University Hospital, and Taipei Veterans General Hospital models, respectively. The explainable predictions were displayed with highlighted input words via the label attention architecture.

Conclusions

Federated learning was used to train the ICD-10 classification model on multicenter clinical text while protecting data privacy. The model’s performance was better than that of models that were trained locally.

Automatic International Classification of Diseases Coding System: Deep Contextualized Language Model With Rule-Based Approaches

BACKGROUND

The tenth revision of the International Classification of Diseases (ICD-10) is widely used for epidemiological research and health management. The clinical modification (CM) and procedure coding system (PCS) of ICD-10 were developed to describe more clinical details with increasing diagnosis and procedure codes and applied in disease-related groups for reimbursement. The expansion of codes made the coding time-consuming and less accurate. The state-of-the-art model using deep contextual word embeddings was used for automatic multilabel text classification of ICD-10. In addition to input discharge diagnoses (DD), the performance can be improved by appropriate preprocessing methods for the text from other document types, such as medical history, comorbidity and complication, surgical method, and special examination.

OBJECTIVE

This study aims to establish a contextual language model with rule-based preprocessing methods to develop the model for ICD-10 multilabel classification.

METHODS

We retrieved electronic health records from a medical center. We first compared different word embedding methods. Second, we compared the preprocessing methods using the best-performing embeddings. We compared biomedical bidirectional encoder representations from transformers (BioBERT), clinical generalized autoregressive pretraining for language understanding (Clinical XLNet), label tree-based attention-aware deep model for high-performance extreme multilabel text classification (AttentionXLM), and word-to-vector (Word2Vec) to predict ICD-10-CM. To compare different preprocessing methods for ICD-10-CM, we included DD, medical history, and comorbidity and complication as inputs. We compared the performance of ICD-10-CM prediction using different preprocesses, including definition training, external cause code removal, number conversion, and combination code filtering. For the ICD-10 PCS, the model was trained using different combinations of DD, surgical method, and key words of special examination. The micro F₁ score and the micro area under the receiver operating characteristic curve were used to compare the model's performance with that of different preprocessing methods.

RESULTS

BioBERT had an F₁ score of 0.701 and outperformed other models such as Clinical XLNet, AttentionXLM, and Word2Vec. For the ICD-10-CM, the model had an F₁ score that significantly increased from 0.749 (95% CI 0.744-0.753) to 0.769 (95% CI 0.764-0.773) with the ICD-10 definition training, external cause code removal, number conversion, and combination code filter. For the ICD-10-PCS, the model had an F₁ score that significantly increased from 0.670 (95% CI 0.663-0.678) to 0.726 (95% CI 0.719-0.732) with a combination of discharge diagnoses, surgical methods, and key words of special examination. With our preprocessing methods, the model had the highest area under the receiver operating characteristic curve of 0.853 (95% CI 0.849-0.855) and 0.831 (95% CI 0.827-0.834) for ICD-10-CM and ICD-10-PCS, respectively.

CONCLUSIONS

The performance of our model with the pretrained contextualized language model and rule-based preprocessing method is better than that of the state-of-the-art model for ICD-10-CM or ICD-10-PCS. This study highlights the importance of rule-based preprocessing methods based on coder coding rules.

Prevalence of Vitamin D Deficiency and Associated Factors in Critically Ill Patients: A Multicenter Observational Study

Background: Vitamin D deficiency is common in the general population worldwide, and the prevalence and severity of vitamin D deficiency increase in critically ill patients. The prevalence of vitamin D deficiency in a community-based cohort in Northern Taiwan was 22.4%. This multicenter cohort study investigated the prevalence of vitamin D deficiency and associated factors in critically ill patients in Northern Taiwan. Methods: Critically ill patients were enrolled and divided into five groups according to their length of stay at intensive care units (ICUs) during enrolment as follows: group 1, <2 days with expected short ICU stay; group 2, <2 days with expected long ICU stay; group 3, 3-7 days; group 4, 8-14 days; and group 5, 15-28 days. Vitamin D deficiency was defined as a serum 25-hydroxyvitamin D (25(OH)D) level < 20 ng/ml, and severe vitamin D deficiency was defined as a 25(OH)D level < 12 ng/ml. The primary analysis was the prevalence of vitamin D deficiency. The exploratory analyses were serial follow-up vitamin D levels in group 2, associated factors for vitamin D deficiency, and the effect of vitamin D deficiency on clinical outcomes in critically ill patients. Results: The prevalence of vitamin D deficiency was 59% [95% confidence interval (CI) 55-62%], and the prevalence of severe vitamin D deficiency was 18% (95% CI 15-21%). The median vitamin D level for all enrolled critically ill patients was 18.3 (13.7-23.9) ng/ml. In group 2, the median vitamin D levels were <20 ng/ml during the serial follow-up. According to the multivariable analysis, young age, female gender, low albumin level, high parathyroid hormone (PTH) level, and high sequential organ failure assessment (SOFA) score were significantly associated risk factors for vitamin D deficiency. Patients with vitamin D deficiency had longer ventilator use duration and length of ICU stay. However, the 28- and 90-day mortality rate were not associated with vitamin D deficiency. Conclusions: This study demonstrated that the prevalence of vitamin D deficiency is high in critically ill patients. Age, gender, albumin level, PTH level, and SOFA score were significantly associated with vitamin D deficiency in these patients.

Factors associated with critical care nurses' acute stress disorder after patient death

BACKGROUND

In intensive care units, patient death can have a negative psychological influence on the patient's nurse. However, how the frequency of events and factors contributed to acute stress among nurses remains unknown.

OBJECTIVE

The objective of this study was to explore the prevalence of and the factors affecting acute stress disorder among intensive care unit nurses after their patient death.

METHODS

Nurses from five adult intensive care units whose patient had died during the nurses' working shift were recruited from July 2018 to April 2019. Bryant's Acute Stress Disorder Scale, the Beck Anxiety Inventory, and the Beck Depression Inventory-II were used to measure acute stress, depression, and anxiety. Descriptive statistics, chi-square tests, independent sample t-tests, and stepwise logistic regression were used for data analysis.

RESULTS

In total, 119 nurses were enrolled. Nearly one in three nurses (29.4%) had suffered from acute stress disorder after their patient had died. Nurses experienced a higher risk of acute stress disorder when their patients underwent cardiopulmonary resuscitation before death (odds ratio [OR] = 13.75, 95% confidence interval [CI]: 2.59-72.95), when their patients died unexpectedly (OR = 4.88, 95% CI: 1.16-20.56), and when they experienced verbal abuse from the patient family at the patient death (OR = 4.61, 95% CI: 1.18-18.05) compared with their counterparts.

CONCLUSION

Intensive care unit nurses often experience acute stress disorder after their patient death. The nurses of patients who underwent cardiopulmonary resuscitation before death and/or who died unexpectedly and/or nurses who were subjected to verbal abuse by the patient's family were at higher risk of acute stress disorder. A comprehensive program aimed at improving the knowledge, skills, and resilience of nurses is needed.

Fisetin Induces Apoptosis of HSC3 Human Oral Cancer Cells Through Endoplasmic Reticulum Stress and Dysfunction of Mitochondria-mediated Signaling Pathways

BACKGROUND/AIM

Oral cancer has been reported to be one of the major cancer-related diseases in human populations and the treatment of oral cancer is still unsatisfied. Fisetin, is a flavonoid from plants and has several biological activities such as antioxidant, anti-inflammatory and anticancer function, but its cytotoxicity in human oral cancer cells is unknown. In the present study, we investigated fisetin-induced cytotoxic effects on HSC3 human oral cancer cells in vitro. Materials and Methods/Results: We used flow cytometric assay to show fisetin induced apoptotic cell death through increased reactive oxygen species and Ca²⁺, but reduced the mitochondrial membrane potential and increased caspase-8, -9 and -3 activities in HSC3 cells. Furthermore, we also used 4' 6-diamidino-2-phenylindole staining to show that fisetin induced chromatin condensation (apoptotic cell death), and Comet assay to show that fisetin induced DNA damage in HSC3 cells. Western blotting was used to examine the levels of apoptotic-associated protein and results indicated that fisetin increased expression of pro-apoptotic proteins such as B-cell lymphoma 2 (BCL2) antagonist/killer (BAK) and BCL2-associated X (BAX) but reduced that of anti-apoptotic protein such as BCL2 and BCL-x, and increased the cleaved forms of caspase-3, -8 and -9, and cytochrome c, apoptosis-inducing factor (AIF) and endonuclease G (ENDO G) in HSC3 cells. Confocal microscopy showed that fisetin increased the release of cytochrome c, AIF and ENDO G from mitochondria into the cytoplasm.

CONCLUSION

Based on these observations, we suggest that fisetin induces apoptotic cell death through endoplasmic reticulum stress- and mitochondria-dependent pathways.

Evaluation of New Biphasic Calcium Phosphate Bone Substitute: Rabbit Femur Defect Model and Preliminary Clinical Results

Autogenous bone grafting, used to repair bone defects, is limited and the donor site can experience complications. Compared to autogenous bone graft, artificial bones have different porosity, which might make them suitable alternatives to bone grafts. Here, two porous biphasic calcium phosphate bone substitutes, namely Bicera™ and Triosite™, are used in an animal study and clinical practice to find a suitable porosity for implantation. Bicera™ and Triosite™ consist of 60 wt% hydroxyapatite and 40 wt% β-tricalcium phosphate, with the porosity of Bicera™ (82%) being higher than that of Triosite™ (70%). In the animal study, the implantation procedure was carried out on twenty-four female New Zealand rabbits. 12 weeks after implantation, the new bones were well infiltrated into the Bicera™ and Triosite™ bone grafts. In the clinical study, patients with comminuted fracture, fracture nonunion, or arthrodesis were included in the study of bone substitution with Bicera™. 27 patients underwent fracture fixation treatment. Bone healing of 22.22% (6/27) of patients happened within 3 months after the surgery, and that of 66.67% (18/27) of patients happened within 6 months. These results reveal that Bicera™ has good incorporation with host bone, and that new bone is able to grow within the porous structure, giving it high potential in the treatment of bone defects.

Effects of diallyl trisulfide on induction of apoptotic death in murine leukemia WEHI-3 cells in vitro and alterations of the immune responses in normal and leukemic mice in vivo

Diallyl trisulfide (DATS), a chemopreventive dietary constituent and extracted from garlic, has been shown to against cultured many types of human cancer cell liens but the fate of apoptosis in murine leukemia cells in vitro and immune responses in leukemic mice remain elusive. Herein, we clarified the actions of DATS on growth inhibition of murine leukemia WEHI-3 cells in vitro and used WEHI-3 cells to generate leukemic mice in vivo, following to investigate the effects of DATS in animal model. In in vitro study, DATS induced apoptosis of WEHI-3 cells through the G0/G1 phase arrest and induction of caspase-3 activation. In in vivo study DATS decreased the weight of spleen of leukemia mice but did not affect the spleen weight of normal mice. DATS promoted the immune responses such as promotions of the macrophage phagocytosis and NK cell activities in WEHI-3 leukemic and normal mice. However, DATS only promotes NK cell activities in normal mice. DATS increases the surface markers of CD11b and Mac-3 in leukemia mice but only promoted CD3 in normal mice. In conclusion, the present study indicates that DATS induces cell death through induction of apoptosis in mice leukemia WHEI-3 cells. DATS also promotes immune responses in leukemia and normal mice in vivo.

Phenethyl isothiocyanate inhibits in vivo growth of subcutaneous xenograft tumors of human malignant melanoma A375.S2 cells

Numerous studies have shown that phenethyl isothiocyanate (PEITC) induces apoptosis of different types of human cancer cell lines, however, there are no reports showing that PEITC inhibits tumor growth in a xenograft model of melanoma in nude mice. We investigated effects of PEITC on the growth of xenografted A375.S2 cell tumors in nude BALB/c mice. A375.S2 cancer cells were inoculated subcutaneously into the lower flanks of mice. Seven days post-inoculation, mice having one palpable tumor were randomly divided into three groups and injected intraperitoneally with PEITC (0, 20 and 40 mg/kg). PEITC reduced tumor weight but total body weight was unaffected. These in vivo results provide support for further investigations to determine the potential use of PEITC as an anticancer drug.

Effects of chitosan on xenograft models of melanoma in C57BL/6 mice and hepatoma formation in SCID mice

According to the World Health Organization, Complementary and alternative medicine (CAM) is a comprehensive term referring to traditional medical treatments and various forms of indigenous medicines, also known as indigenous or folk medicine. Cancer patients often use CAM in the form of nutritional supplements, psychological techniques and natural medical approaches in the place of or in parallel to conventional medicine. The present study aimed to determine if Chitosan can inhibit lung metastasis and hepatoma formation, by studying xenograft of B16F10 melanoma cells in C57BL/6 mice and of Smmu 7721 cells in SCID mice, respectively. For the lung metastasis model, after a five-week treatment, the survival rates of B6 mice were 15% for the control group and 35%, 20%, 45% and 40% for the 320,000 kDa, 173,000 kDa, 86,000 kDa and 8,000 kDa molecular-weight treatment groups, respectively. Chitosan treatment dramatically increased lifespan and inhibited tumor metastasis especially in treatment groups of the low-molecular weight compound. For the hepatoma growth model, the size of the liver tumor mass was approximately >14 mm in the control group. In comparison to the control group, the tumor mass grew slowly with Chitosan treatment, especially at the low-molecular weight treatment group. Chitosan slowed-down the rate of tumor growth but did not inhibit tumor formation. Data presented herein demonstrate that Chitosan has anticancer effects and thus further study of the substance is warranted to examine for mechanisms of action and optimal dosage.

Triptolide induces S phase arrest via the inhibition of cyclin E and CDC25A and triggers apoptosis via caspase- and mitochondrial-dependent signaling pathways in A375.S2 human melanoma cells

Triptolide (TPL), a diterpene triepoxide compound, extracted from Tripterygium wilfordii Hook F. [a traditional Chinese medicinal herb (TCM)], has demonstrated great chemotherapeutic potential for the treatment of tumors. However, the anticancer mechanisms of action of TPL in human skin cancer remain to be further investigated. In this study, we used A375.S2 human melanoma skin cancer cells as a model to investigate the effect of TPL on cell death. A375.S2 cells were treated with various concentrations of TPL for different periods of time and investigated the effects on cell cycle distribution and apoptosis were investigated. The data showed that TPL induced cell morphological changes, decreased the percentage of viable cells, and induced S phase arrest and apoptosis in A375.S2 cells in a concentration- and time-dependent manner. Furthermore, we used flow cytometry analysis and the data showed that TPL promoted reactive oxygen species, NO and Ca2+ production, decreased the mitochondrial membrane potential (ΔΨm) and increased the activity of caspase-3, -8 and -9 in the A375.S2 cells. Western blot analysis showed that TPL promoted the expression of p21 and p27 but inhibited that of cyclin A and CDC25A, leading to S phase arrest. Furthermore, the data also showed that TPL promoted the expression of Fas and FasL and increased the activity of caspase-3, -8 and -9, cytochrome c, Bax, apoptosis-inducing factor (AIF) and endonuclease G (Endo G); however, the expression of Bax was decreased, leading to apoptosis. Based on these observations, TPL induces apoptosis in A375.S2 cells through Fas-, caspase- and mitochondrial-mediated pathways.

Butylated hydroxyanisole affects immunomodulation and promotes macrophage phagocytosis in normal BALB/c mice

Butylated hydroxyanisole (BHA), a synthetic antioxidant, has been used in fat and fatty foods to prevent oxidative deterioration. However, the functions of BHA on immune responses in normal mice remain elusive. The aim of the present study was to investigate the effects of oral treatment of BHA on immune responses in normal mice in vivo. BALB/c mice received various treatments. Blood samples were collected and analyzed. Flow cytometry was used to determine the levels of the cell markers. Results showed that BHA did not significantly affect the weight of the animal body and spleen in normal mice. BHA promoted macrophage phagocytosis from peripheral blood mononuclear cells, but did not alter this process in the peritoneal cavity. Furthermore, BHA did not influence natural-killer cell cytotoxicity in normal mice. Notably, BHA promoted the levels of CD3 (T cells) and decreased the level of CD19 (B cells), but did not significantly affect the levels of CD11b (monocytes) and macrophages (Mac-3) in normal mice. Based on these observations it can be concluded that BHA promotes immune responses by increasing T cells and activating phagocytosis by macrophages in normal mice. However, the molecular mechanisms require further investigation.

Acute hemorrhagic rectal ulcer: an important cause of lower gastrointestinal bleeding in the critically ill patients

BACKGROUND AND AIM

The occurrence of acute hemorrhagic rectal ulcer (AHRU) in patients in the intensive care unit (ICU) has not been well investigated. The aims of this study were to evaluate the clinical manifestations and outcomes in these patients.

METHOD

The patients developing significant acute lower gastrointestinal (LGI) bleeding after ICU admission from July 2002 to December 2007 were retrospectively reviewed. Bedside colonoscopy was performed within 24 h of bleeding, and those patients with bleeding from AHRU were studied. Ulcers with stigmata of recent bleeding were treated with endoscopic hemostasis, and the outcome of these patients was analyzed.

RESULTS

AHRU occurred in 36 of 114 patients (31.6%) and was the most common cause of acute LGI bleeding after ICU admission. Most patients had comorbidities, such as respiratory failure, renal failure, diabetes mellitus, or atherosclerosis. Fourteen patients (38.9%) developed hypovolemic shock after the onset of bleeding. Endoscopic therapy was performed in 29 patients with 97.2% success rate for hemostasis. Fourteen patients (48.3%) developed rebleeding within 4 weeks. This was controlled by repeated endoscopic intervention. Anticoagulant use was the risk factor for rebleeding after treatment. The survival rate at 4 weeks after bleeding was 52.8%. Logistic regression analysis revealed thrombocytopenia (platelet count <150,000/mm(3)) and more than one comorbidity were independent predictors for mortality.

CONCLUSIONS

AHRU is an important etiology of acute LGI bleeding in the patients with critical illness. Bedside colonoscopy is helpful for early diagnosis and treatment. The underlying comorbidities of the patients influence the outcome after bleeding.

Danthron inhibits murine WEHI-3 cells in vivo, and enhances macrophage phagocytosis and natural killer cell cytotoxic activity in leukemic mice

Danthron has been shown to induce apoptotic cell death, and inhibit migration and invasion of human gastric or brain cancer cells in vitro. However, there is no report addressing whether danthron affects murine leukemia cells or immune responses in vivo. Herein, this study focused on the in-vivo effects of danthron on WEHI-3 leukemia in mice and immune responses in vivo. The results indicated that danthron reduced spleen weight and increased the percentage of cells with CD3 and CD19 markers, indicating that differentiation of the precursors of T- and B-cells was promoted in the leukemic mice. The results also showed that danthron promoted the activity of phagocytosis by macrophages isolated from the peritoneal cavity but had no effect on peripheral blood mononuclear cells. Danthron also promoted natural killer cell cytocytic activity at an effector and target cell ratio of 100:1 in comparison with leukemic animals in vivo. Taken together, these results demonstrated that application of danthron might affect WEHI-3 leukemia in mice and modulate immune responses in vivo.

Arterial misplacement of a femoral central venous catheter complicated with acute arterial occlusion

Femoral vein catheterization is often carried out during resuscitation and in critical care units. It is usually achieved via a blind, external landmark-guided technique, through manual localization of the femoral artery. However, this approach can be challenging in patients with severe shock because of absence or ambiguity of the arterial pulse. We report a case of inadvertent cannulation, with a large-bore catheter, of the right femoral artery, which was mistaken as a venous route for medication and massive transfusion. The large cannula caused direct mechanical obstruction, while intra-arterial medications induced vascular injury and vasospasm. Both factors led to thrombosis and occlusion of the right external iliac artery, thus jeopardizing the distal blood supply, and ultimately resulting in muscle necrosis of the involved limb, and the need for thrombectomy and extensive fasciotomy to salvage the extremity. This case highlights that correct placement of a central venous catheter should be clearly ascertained before the catheter is used for medical treatment, especially when catheterization is performed in shock status.

Immunohistochemical localization of choline acetyltransferase in rabbit forebrain

uinea pig antiserum specific to the purified bovine choline acetyltransferase was used to demonstrate the localization of this enzyme in rabbit forebrain by the peroxidase-antiperoxidase immunohistochemical method. Choline acetyltransferase was localized in olfactory bulb, olfactory tract, olfactory tubercle, piriform cortex, septum, diagonal band, basal ganglia, thalamus, hypothalamus, subthalamus, habenula, cerebral cortex, hippocampal region, corpus callosum, internal capsule, fornix, longitudinal striae and other areas. The findings reflect the distribution of cholinergic axons and, possibly, their terminals. These observations correlate well with biochemical determinations of choline acetyltransferase and with previously proposed cholinergic pathways.