Thirty-eight patients with chronic respiratory failure, including 28 patients receiving HOT, were studied from the psychosomatic viewpoint. Assessment of psychological state was based on the following 5 psychometrical tests; Cornell Medical Index (CMI), Self Rating Questionnaire for Depression (SRQ-D), State Trait Anxiety Inventory (STAI), Minnesota Multiphasic Personality Inventory Alexithymia Scale (MMPI-Alexithmia scale) and egogram Check List (ECL). Tweleve HOT patients were also interviewed individually over a three-year period. The ECL showed overadaptation of non-HOT female patients, as compared with their HOT counterparts. There was an inverse correlation between PaO2 (room air) and the SRQ-D score. A significant correlation was also found between age and the SRQ-D score in both HOT and non-HOT patients. HOT patients were observed for 3 years (1-4 years after HOT introduction). Depressive state was observed in 16.6-36.4% of the patients throughout this observation period. Alexithymia was noted in 25.0% only 1 year after the initiation of HOT. However, the average scores for SRQ-D, STA-I and the MMPI-Alexithymia scale remained unchanged for 3 years. PaCO2 (room air, O2, inhalation) was significantly related to the STAI-1 score up to 2 years after starting HOT. Furthermore, a significant correlation was seen between HOT duration and MMPI-Alexithymia scales in HOT patients. Blood gas analysis parameters (delta PaO2, PaCO2 and pH) correlated significantly with respiratory scores on CMI somatic profiles. These results suggest that blood gas status (PaO2, PaCO2) may have an effect on the psychological states of patients early in the course of HOT. However, social factors, such as familial or economical problems, seem to have a greater influence in later stages. Thus, more attention must be given to psychosomatic treatment in the care of HOT patients.