Eur Respir J 1988, 1, 952- 958

A national register for long-term oxygen therapy in chronic hypoxia: preliminary results

K. Strom, J. Boe

A national register for long-term oxygen therapy in chronic hypoxia: prelimi· Division of Chest Medicine, Central Hospital, nary results. K. Strom, J. Boe. Vaxjo, . ABSTRACT: A national register of patients undergoing long-term oxygen The Swedish Society of Chest Medicine. therapy in Sweden was started In 1987. Of tbe population of 8.4 miUion, All Depa11ment.s of Chest Medicine, Sweden. 560 patients (267 males) were registered as undergoing domiciliary oxy· gen treatment on January 1, 1987. These registered patients, aged between Keywords: Chronic hypoxaemia; domiciliary 2-86 yrs (mean age 65 yrs), constituted some 90% of aU patients receiv­ oxygen; national register. ing domiciliary oxygen therapy because of chronic hypoxaemia on that Received: February, 1988; Accepted after revi· date. The chronic respiratory diseases leading to hypoxaemla (more than sion July 3, 1988. one diagnosis could be registered for each patient) were: chronic obstruc­ tive pulmonary disease (COPD) (393), sequelae from pulmonary tubercu­ The register is supported by grants from the losis (92), thoracic deformity (97), Interstitial fibrosis (44), benign pleural Swedish Association against Heart and Chest disease (22) and others (84). Concentrators were used by 253 patients and Diseases, the Kronoberg County Council and the high-pressure compressed gas cyl.lnders by 307. Arterial blood gas analy­ National Association for the Prevention of ses were registered for 547 patients. The mean arterial oxygen tension Asthma and Allergies. ) (Pao1 when breathing air was 6.5±1.1 kPa and 9.0±1.4 kPa when breath· lng oxygen. The mean arterial carbon dioxide tension (Paco ) without 1 oxygen was 6.5±1.5 kPa and 6.8±1.5 kPa with oxygen. The register forms a data-base whlcb can be used for the evaluation of different home oxy­ gen systems, regional differences In the access to treatment and treatment performance and decisions relating to health care economics. Eur Respir J., 1988, I, 952-958.

It has been suggested for some twenty years that High-pressure compressed gas cylinders and station­ long-term oxygen therapy may improve pulmonary ary oxygen concentrators are available for domestic use. hypertension [1, 2}. The controlled Nocturnal Oxygen Ambulatory patients on continuous oxygen also use Therapy Trial (NOIT) (1980) [3, 4) and Medical small gas cylinders. Only compressed gas cylinders are Research Council (MRC) trial (1981) [5] revealed paid for by the National Insurance Service. Concentra­ improved survival in the case of chronic respiratory tors are paid for by some county health authorities. insufficiency caused by chronic obstructive pulmonary Due to the regional differences in access to and the disease (COPD) as a result of long-term domiciliary performance of domiciliary oxygen therapy the Swed­ oxygen therapy. The MRC study showed that 15 h of ish Society of Chest Medicine started a register of oxygen, including night-time, improved survival when patients undergoing long-term home oxygen therapy. compared with no oxygen. The NOIT study demon­ This register forms a data-base that can be used to strated improved survival and neuropsychiatric function evaluate the long-term effects of oxygen therapy at in patients receiving "continuous" oxygen (slightly less home in patients with chronic hypoxia of varying ori­ than 20 h) compared with patients receiving only 12 h gin. It also makes it possible to compare different home of oxygen, including their hours of sleep. No controlled oxygen systems and could be used to compare differ­ clinical trials have been conducted to evaluate the ent health regions in terms of home oxygen therapy. effect of long-term home oxygen therapy in diseases This register could also guide health authorities when causing chronic hypoxaemia and cor pulmonale other it comes to the reallocation of resources for home oxy­ than COPD. Prescription principles similar to those used gen therapy. This is the first presentation of the for COPD have been recommended for these diseases register which includes all the registered patients under­ [6]. going domiciliary oxygen therapy on January I, 1987. Sweden with its 8.4 million inhabitants, is divided into seven health regions, each consisting of 2- 5 coun­ ties. The use of domiciliary oxygen therapy has spread Methods considerably following the NOIT and MRC studies and is usually prescribed by chest physicians, except in All the Departments of Chest Medicine in Sweden are some areas without a county chest department. participating in the study. One to four physicians are THE SWEDISH OXYGEN REGISTER 953

responsible for the registration of data in each county. hypoxic patients who were not registered lived in Stock­ The record forms were discussed in a reference group holm. An additional check on the use of home oxygen and contain information on age, sex, diagnosis, present in was therefore made by the National or former occupation, smoking habits, medication, pres­ Corporation of Swedish Phannacies and the results were ence of peripheral oedema, blood gas analyses and para­ reported to the registration centre in Vl:1xjO. meters linked to oxygen treatment (such as the form of % of patients admirustration, type of nasal cannulae, oxygen dose, 25 ----- Females length of treatment in hours per day). --Males Table 1. - Prescribing guidelines for long-term oxygen therapy in chronic hypoxaemia 20

Patient selection criteria: 15 - COPD or other diseases leading to chronic hypoxaemia; - stable course of disease on optimum medical therapy for at least three weeks; 10 - arterial oxygen tension when breathing air still less than 1.0-1.5 kPa or signs of cor pulmonale or haematocrit >50% and room air of around 7.5 5 Pao2 kPa. Oxygen dose:

- oxygen treatment should be started and the dose adjusted <40 45 50 55 60 65 70 75 80 85 90 while the patient is in hospital; Fig. 1. - Age and sex distribution (n=560). ••• Females; -Males - continuous flow by double or single nasal carmulae, mini­ mum 16 h. (hours of night included), preferably 24 h; This survey found 76 unregistered patients receiving

- arterial Pao2 of 8 kPa or higher. domiciliary oxygen therapy because of respiratory faiJ­ ure, including patients who had been asked to par­ : COPD: chronic obstructive pulmonary disease; Pao2 arterial ticipate in the register but had refused. In an additional oxygen tension. seven patients no diagnosis was found. Taken together, these reports have shown that some 90% (1987-1988) The guidelines for prescribing long-term domiciliary of all patients in Sweden receiving domiciliary oxygen oxygen therapy have been issued by the Swedish therapy because of respiratory failure were included in Association of Chest Medicine (table I). Patients al­ ll1e register. There were 2fj7 men and 293 women with ready on home oxygen therapy on January 1, 1987, and a mean age of 65 yrs (range 2- 86 yrs). Two patients patients discharged from hospitals on home oxygen for were less than twenty years old (fig. 1). A hisLOry of the ftrst time after this date are included in a prospec­ peripheral oedema was found in 327 patients (63%). tive study. All the record forms, including follow-up Twenty-siJt patients (5%) were still at work (housework record forms every six months, are sent to the not included). 507 patients (91 %) lived at home, while registration centre in Vl1xj~ for data processing and sta­ 40 patients livcil in nats or homes for the elderly. tistical evaluation. The data processing is performed Dormciliary oxygen is not recommended for smokers. with the aid of the Quest data-base programme [7). Nevertheless. 44 patients (8%) were noted as current Comparisons between groups were performed using t­ smokers. tests. Mann-Whit:ney non-paramelric tests were used for small groups [8]. Informed consent was obtained from each patient and the study was approved by all ll1e re­ Diagnosis gional Ethical Committees, the Swedish Board of Health More than one diagnosis could be registered for each and Welfare and the Data Inspection Board. patient and the record forms did not permit the rank­ ing of the different diagnoses noted. COPD was one Results cause of hypoxaemia in 393 patients; in other words, COPD was not present in 167 patients (30%). Sequelae Patients of tuberculosis was the cause of or a contributory factor to hypoxaemia in 102 patients (18%). In 106 During the first months of 1987, 560 patients under­ patients (19%) neither COPD nor sequelae of tubercu­ going home oxygen treatment because of chronic losis was a factor contributing to respiratory failure hypoxia on January 1, 1987, had been reponed to the (table 2). registration centre in Vbj(}. The prevalence of patients receiving home oxygen therapy due to respiratory fail­ ure was estimated with the aid of reports given by the Arterial blood gas analyses local pharmacies supplying oxygen cylinders and local hospital technical workshops supplying oxygen equip­ Arterial blood gas analyses were registered for 547 ment to the participating physicians. The majority of patients. The most recent blood gas analyses breathing 954 K. STR6M, J. BOE

Table 2. -Cause of hypoxaemia % of patients 30 ,.... --Oxygen No. of I ----- Air patients 25 I n=560 % / 20 I I COPD 393 70 I 15 I Sequelae of tuberculosis 102 18 I ' I pubnonary sequelae 92 16 I. 10 I thoracic deformity 57 10 I I Interstitial fibrosis 44 8 I 5 I Kyphoscoliosis 40 7 Benign pleural disease 22 4 Pulmonary emboli 14 3 <3 4 5 6 7 8 9 10 11 12 13 Sarcoidosis 10 2 Paco2 Neuromuscular diseases 9 2 Fig. 3. - Distribution of arterial carbon dioxide tension (Paco ) on Primary pulmonary hypertension 7 2 1 room air and oxygen (kPa) (n=547). --- Air;- 0 • Tumour in lung and/or pleura 7 1 2 Pneumoconiosis 7 1 Table 3. - List of medication used by patients with and Miscellaneous 30 5 without COPD

More than one diagnosis could be registered for each patienL Medication No. of No. of COPD: chronic obstructive pulmonary disease. patients patients with without %of patients COPD COPD 40 --Oxygen n=393 % n=167 % I \ ----- Air I \ " Oral B agonist 268 68 47 28 30 I \ 2 I \ Inhaled fl -agonist 311 79 54 32 I \ 2 I \. Theophylline 292 74 59 35 I \ 20 I Oral steroid 192 49 38 23 I I Inhaled steroid 132 34 16 10 I Digoxin/digitoxin 134 34 65 39 10 I I ' Loop diuretic 329 84 139 83 I /' Spironolactone 145 37 61 37 Calcium antagonist 26 7 17 10 ... -"'" 0 Anticoagulant 10 3 10 6 <3 4 5 6 7 8 9 10 11 12 13 14 15 PaO:?.kPa COPD: chronic obstructive pulmonary disease.

) Fig. 2. - Distribution of arterial oxygen tension (Pao2 on room air and oxygen (kPa) (n=547). --- Air; -02 . patients with COPD, as was expected, but they were also used by approximately one third of the patients room air revealed hypoxaemia (arterial oxygen tension without COPD. Loop diuretics were used by 468 pa­ (Paoz> of less than 8 kPa) .in 91% .of the 547 p~tie~ts tients (84%) (table 3). (497 patients). Hypercapma (artenal carbon dtoxtde tension (Paco) of 6.1 kPa or higher) breathing air was 2 Oxygen therapy found in 59% of the 547 patients and in 15% (82 pa­ tients) the Paco was 8 kPa or higher. When breathing 2 Oxygen concentrators from four different manufactur­ oxygen the mean Pao increased from 6.5±1.1 to 2 ers (102 Briox, 85 OECO, 35 Zefir and 31 DeVilbiss) 9.0±1.4 kPa (fig. 2), while the mean Paco increased 2 were used by 253 patients (45%). Thirteen patients also from 6.5±1.5 to 6.8±1.5 kPa (fig. 3). With oxygen the received home ventilator treatment. Double nasal can­ i.e. Pao2 exceeded 8 kPa in 441 patients, 80% of the nulae were used by 412 patients (76%) and single 547 patients in whom arterial blood gas analyses were nasal cannulae by 120 patients (22%). Six patients registered. No difference was found between the Pa0 2 had transtracheal catheters. The humidification of and PaC0 on room air and oxygen in COPD patients 2 oxygen was common, regardless of the oxygen dose. and patients with other diagnoses. 1 The oxygen dose varied between 0.2-8 l·min· , mean 1.4 /·min·1 (fig.4). No difference in oxygen dose was Medication found between COPD patients and patients without COPD. Oxygen was prescribed for a mean of 19 h Bet3z-agonists and theophylline were used by most (range 1-24 h) (fig. 5). THE SWEDISH OXYGEN REGISTER 955

% of patients Table 5. - Mean duration of oxygen therapy (months) 30 since January 1, 1987, related to diagnosis 25 Diagnosis Mean therapy No. of 20 duration patients

15 COPD 21 385 Sequelae of tuberculosis 29 99 10 pulmonary 27 89 5 thoracic deformity 35 53 Others (COPD, sequelae of tuberculosis excluded 24 102 0·0.5 0.5· 1 1-1.5 1.5-2 2·2.5 2.5-3 3·3.5 3.5-4 >4 Interstitial fibrosis 19 43 Litres/min Kyphoscoliosis 40 38 Benign pleural disease 30 21 Fig. 4. - Oxygen flow rate (n=559). Pulmonary emboli 29 13 Sarcoidosis 25 10 Table 4. -Age at start of home oxygen therapy (years, Neuromuscular diseases 25 7 mean values, range) Primary pulmonary hypertension 22 7

Cause of Age )'IS COPD: chronic obstructive pulmonary disease. hypoxaemia Males Females Missing data

Patients/million > 45 yr COPD 66 (26-M) 62 (344!3) n=393 n=195 n=190 8 400 c:J Not COPD Sequelae of 67 (45- 78) 65 (46-81) 1?221 COPD tuberculosis n=43 n=56 3 Others 56 (2-77) 57 (33-84) n=106 n=43 n=59 4 All causes 64 (2- 84) 61 (33-84) n=560 n=262 n=283 15

COPD: chronic obstructive pulmonary disease.

Age ar start of oxygen therapy and mean duration of oxygen therapy. The age of the patients at the start of Sth Lin Lund Gbg Upp Ore Umea home oxygen therapy varied depending on th eir sex and Health Region the cause of hypoxaemia (table 4). Among COPD pa­ tients women were younger when therapy began Fig. 6. - Regional distribution of patients aged >45 yrs per mUlion inhabitants aged >45 yr-s. Patients with COPD (n,.382) and pal ients wi.thout COPD (n=l58). l:lcalth regions: Sth: Stockholm· Un: Linkoping; Lund; Gbg: ; Upp: Uppsllla; 6re: 6rcbro; UmeA.; COPD: chronic obstructive pulmonary disease. % of patients c:=J N01. COPO; ~COPD. 50 Table 6. - Mean duration of oxygen therapy (months) since January 1, 1987, related to diagnosis and sex 40 Diagnosis Mean therapy No. of 30 duration patients

20 Males Females Males Females

10 COPD 19 22 195 190

Sequelae of tuberculosis 23 33 43 56 <12 12-15 16 -19 20 - 24 Others (COPD, sequelae Hours/day of tuberculosis excluded) 16 29 43 59 Fig. 5. - Hours of supplementary oxygen per day as prescribed for patients (n=560). COPD: chronic obstructive pulmonary disease. 956 K. STROM, I. BOE

(p

the two countries was the proportion of oxygen Sven ss~n) , Linkoping (A. Brun din), Lund (B. Midgrcn), Malmo (E. Pihulainen), Skovde (f. prescriptions made by chest physicians: 44% in France MAnsson), Stochholm (E. Ripe, C-1!. Ericsson, E. lhre, and more than 80% in Sweden. This may have some­ H. Heilbom). (H. LOwgreo), UddevaJJa (L­ thing to do with the fact that only 59.4% of the French G. Carlsson), UmcA (L. Rosenhall), (G. Boman), (Z. Varga), Vlinersborg (L. Grettve, B. patients were prescribed oxygen for more than 12 h Ostholm), Vasterls (S. SOrenson), Orcbro (M. Ahoo), daily. Oxygen treatment for Jess than IS h daily is of Ostcrsund (P.O. Rydstrilm). We arc also indebted to E. Saner.~trom , Vbjo and L. Gustafsson, Dcpanment unproven therapeutic benefit, at Jeast in COPD patients, o( SUitistics, University of Ume!, for data processing and only 8% of the registered patients received so little and SUIIistical evaluation, and 10 I!.B. Nilsson for treatment secretarial assistance. The geographicaJ distribution of home oxygen therapy was found to be markedly uneven. The same References situation has been found in England where home oxygen therapy is usually prescribed by general 1. Levine BE, Bigelow DB. Hamsl.ra RO et at. - The role practitioners [22, 23]. During this century the prevaJence of long-term continuous oxygen admlnisLTation in patients with of tuberculosis has been higher in the north of Sweden chronic airway obstruction with hypoxacm.ia. A rv1 ltllem Med, [24] and this may account for a small proportion of the 1967. 66, 639-(i50. rcgionaJ differences. The large geographicaJ variation in 2. Abraham AS, Cole RB, Bishop JM. - Reversal of pul­ monary hypertension by prolonged oxygen adminisLTation to home oxygen therapy for COPD cannot be explained patients with chronic bronchitis. Circ Res. 1968, 23, 147- 157. by variations in the age of the popuJation or in the 3. Nocturnal Oxygen Therapy Trial Group. - Continuous or prevalence of the disease estimated from the mean an­ nocturnal oxygen therapy in hypoxaemic chronic obstructive nual crude death rate from bronchitis, emphysema and lung disease. Ann lmern Med, 1980, 93, 391-398. asthma. The mean age at the start of oxygen therapy 4. Heaton RK, Grant I, McSweeny J, Adams K. Petty TL. as well as the mean duration of therapy was the same - The psychologic effects of continuous and nocturnal oxy­ in the county with the highest prevalence of home oxy­ gen therapy in hypoxaem.ic chronic obstructive pulmonary gen therapy as in the whole of Sweden. This suggests disease. Arch ltllem Med, 1983, 143, 1941- 1947. that the start of oxygen therapy at an earlier stage in 5. Report of the Med.ical Research Council Working Party. the course of the disease does not explain the greater - Long-tenn domiciliary oxygen therapy in chronic cor pul­ use of oxygen in this county. This county has one of monale complicating chronic bronchitis and emphysema. Lan­ cet, 1981, i, 681-686. the major university chest clinics. A substantial part of 6. Petty TL, Flenley DC. - Domiciliary and ambulatory the geographical variation can be explained by vari­ oxygen. In:. Respiratory Medicine 4. Churchill Livingstone, ations in the knowledge and interest of the physicians Edinburgh. 1986. in charge of patients with respiratory failure. In some 7. Gustafsson L. -Quest A programme system for statisti­ counties domiciliary oxygen therapy has a long cal and epidemiological data analysis. University of UmeA, tradition, while in others the therapy has only 1987. been introduced in recent years. With a few exceptions 8. Colton T. - In: Statistics in Medicine. Little, Brown and domiciliary oxygen is used more frequently in counties Co., 1975. with a chest clinic than in countries where patients 9. Causes of death. - Official Statistics of Sweden. National with respiratory failure are seen by specialists in CenLTal Bureau of Statistics, Stockholm, 1985-1987. 10. Cooper CB, Waterhouse J, Howard P. - Twelve year other disciplines. The highest county prevalence of clinical study of patients with hypoxic cor pulmonale given home oxygen therapy was 0.048% of the population long-tenn domiciliary oxygen therapy. Thorax, 1987, 42, aged 45 yrs or over. This county had the second low­ 105-110. est death rate for COPD. lt can be assumed that the 11. Proceedings of the 4th Congress of SEP. - Observatory prevalence of respiratory failure is no lower in the report of France. 8000 patients with home respiratory county as a whole and this means that a minimum of assistance. Eur J Respir Dis, 1985, (Suppl. 146). 1,700 patients may be candidates for home oxygen ther­ 12. Gerard M, Roberl D, Salamand J, Buffat J, Chemorin apy in Sweden. B, Bertoye A. - Etude de la survie chez 150 insuffisants We have found that home oxygen therapy performs respiratoires chroniques LTacMotomises LTaites par la ventila­ tion assistee a domicile. Lyon Med, 1981, 245, 555-558. well and is prescribed for the right patients when it is 13. Bergofsky EH, Turino OM, Fishman AP. - Cardiorespi­ initiated and controlled by chest physicians. The uneven ratory failure in kyphoscoliosis. MediciM (Ballirrwre), 1959, geographical distribution means that patients in several 38, 263-317. counties have poor access to this therapy. The use of 14. Hoeppner VH, Cockroft OW, Oosman JA, Cotton OJ. ­ concentrators, the most convenient therapy for most Nigh t-time ventilation improves respiratory failure in secon­ patients, also varies to a great extent dary kyphoscoliosis. Am Rev Respir Dis, 1984, 129, 240-243. 15. Libby OM, Briscoe WA, Boyce B, Smith JP. - Acute respiratory failure in scoliosis or kyphosis. Am J Med, 1982, Ac:knowledgements: We are indebted to the pan ici­ 73, 532-538. pating physicians in the following Depanments' of 16. Evans TW, Waterhouse J, Howard P. - Clinical experi­ Chest Medicine (panicipating physician ~ Ul brackets): Boden (L·G. Larsson), Bor4s (L U>fgren), E1kilstuna ence with the oxygen concentrator. Br Med J, 1983, 287, (f. Struwe), (L. Hanberger), G!ilve (M. Udden­ 459-461. feldt), Goteborg (K. Pdusson), (J. MusiJ), 17. Lawson KY. Drwrunond MF, Bishop JM.- Costing new Helsingborg (G. Stiksa), Jonkoping (L. Maroti), (f. Arvidsson), (V. Nemcek), services: long-tenn domiciliary oxygen therapy. Lancet, 1981, (N. Fonder-Jorgensen). Kristianstad (G. 1146-1149. 958 K. STROM, J. BOI!

18. Gould GA. Scon W, Hayhurst MD, Flenley DC.- Tech­ oxygenotherapie au long cours a commence en Suede en nical and clinical assessment of oxygen concentrators. Thorax, 1987. Pour une population de 8.4 millions d'habitants, 560 1985, 40, 811-816. patients (dont 267 hommes) ont ete enregistres comme etant 19. Fulmer JD, Snider GL. - ACCP-NHLBI National SOumis a une oxygenotherapie a domicile le 1er janvier 1987. conference on oxygen therapy. Chest, 1984, 86, 234-247. Ces patients enregistres, ages de 2 a 86 ans (age moyen 65 20. Swedish Social Welfare Board - The humidification of ans), representent environ 90% de tous les patients recevant breathing oxygen. Information from the Swedish Social de l'oxygenotherapie a domicile pour hypoxemie chronique a Welfare Board (in Swedish), 1983, 2. cette date. Les maladies chroniques respiratoires conduisant a 21. Exploitation statistique de pres de 7000 demandes l'hypoxemie ont ete les suivantes (plus d'un diagnostic peut d'oxygenotherapie a domicile pour insuffisance respiratoire etre enregistre pour chaque patient): BPCO (393), sequelles chronique grave (I.R.C.G). Revue Medicate de J'Assurance de tuberculose pulmonaire (92), deformations thoraciques (97), Maladie, 1987, 4, 13-18. fibroses interstitielles (44), affections pleurales benignes (22), 22. Williams BT. - Geographical variations in the supply autres (84 ). Les concentrateurs ont ete utilises chez 253 of domiciliary oxygen. Br Med J, 1981, 282, 1941-1943. patients et des ecylindres de gaz comprime a haute pression 23. Williams BT, Meechan DF, Nicholl JP. - Why does chez 307. Les analyses des gaz du sang arteriel ont ete domiciliary oxygen consumption vary from area to area? J R enregistrees chez 547 patients. La pression arterielle moyenne Coil Gen Pract, 1983, 33, 481-483. d'oxygene sous respiration d'air est de 6.5±1 kPa; elle est de

24. Puranen BI. - Tuberculosis. Its occurrence and causes 9.0±1.4 kPa sous respiration d'oxygene. La Paco2 moyenne in Sweden 1750-1980 (in Swedish). University of UmeA, sans oxygene est de 6.5±1.5 kPa et de 6.8±1.5 kPa sous 1984. oxygene. Ce registre constirue une base de donnees utilisable pour !'evaluation des differents systlimes d'oxygenotherapie a Le regi.stre suedoi.s d'oxygbwtMrapie. Un registre national domicile, pour celle des diff~rences r~gionales dans pour l'oxygenothirapie au long cours dans l'hypoxie l'accessibilite au traitement, pour celle des resultats du traite­ chronique. Resultats preliminaires. K. Strom, J. Boe. menl et de decisions en rapport avec I' economie de Ia sante. RESUME: Un registre national de patients sournis a une Eur Respir J., 1988, 1, 952-958.