ARTICLE IN PRESS

Respiratory Medicine (]]]]) ], ]]]–]]]

Survival of patients on home mechanical ventilation:

A nationwide prospective study
Michael Lauba,b, Bengt Midgrena,

a
Department of Respiratory Medicine, University Hospital, 22185 Lund, Sweden
b
Respiratory Centre East, University Hospital, Copenhagen, Denmark

Received 16 July 2006; accepted 3 October 2006

KEYWORDS Summary
Respiratory failure; Home mechanical ventilation (HMV) is increasingly used as a therapeutic option to patients
Home care; with symptomatic chronic hypoventilation. There is, however, a paucity of solid data on
Survival; factors that could affect prognosis in patients on home ventilation. In the present study,
Predictors for death our aim was to study several factors in these patients with potential influence on survival.
We examined 1526 adult patients from a nationwide HMV register to which data had been
reported prospectively for 10 years. The patients constituted a broad diagnostic spectrum
and the primary outcome in this study was death.
We found by far the poorest survival rate in the ALS patients with only 5% alive after 5
years. Among the other patient groups the survival pattern was more uniform and the
scoliosis, polio and Pickwick patients presented the best survival rate, after 5 years being
around 75%. No factors were associated with a greater hazard for death in the ALS
patients; in the non-ALS patients, however, negative predictors for survival were age,
concomitant use of oxygen therapy, tracheostomy ventilation and start of ventilatory
support in an acute clinical setting. Center size or county specific home ventilation
treatment prevalence did not affect survival.
In conclusion, in a large material of patients on HMV we found by far the poorest survival in
the ALS patients. In the non-ALS patients a number of patient-related factors affected
survival, while the size of the treating center or the regional treatment prevalence
did not.
& 2006 Elsevier Ltd. All rights reserved.

Introduction

One important outcome of home mechanical ventilation

(HMV) is survival. It has been suggested that HMV can
Corresponding author. Tel.: +46 46 171 214. prolong survival and improve quality of life in neuromuscular
E-mail address: [email protected] (B. Midgren). and chest wall disorders,1–5 whereas it may reduce

0954-6111/$ - see front matter & 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.rmed.2006.10.007

Please cite this article as: Laub M, Midgren B. Survival of patients on home mechanical ventilation: A nationwide prospective study.
Respiratory Medicine (2006), doi:10.1016/j.rmed.2006.10.007
 ARTICLE IN PRESS
2 M. Laub, B. Midgren

symptoms of nocturnal hypoventilation and breathlessness start, four counties out of the 26 counties in Sweden,
in other situations.6 A consensus that HMV should be offered (henceforward referred to as ‘‘top counties’’), consistently
as a therapeutic option to patients with symptomatic presented a much higher HMV treatment prevalence
chronic hypoventilation appears to prevail in the field compared to the rest.13 In these counties lived 26% of the
although there are few randomized controlled trials to HMV patients. At the time of this analysis the top counties
support this. had a treatment prevalence of 22 per 100.000 in contrast to
Some guidance regarding the usefulness of HMV can be 14 for the rest.
derived from studies comparing survival with respect to The Swedish Data Inspection Board has approved the
diagnosis.7–10 There are however but a few studies compar- register and the Medical Ethics Committee at the University
ing survival as regards patient related factors such as of Lund has approved the study. For statistical analyses we
physiological and clinical data.3,9 The impact of caregiver used the software STATISTICA version 7.1 to generate
related factors such as center size and regional treatment descriptive data and compile regression models and graphs.
prevalence has only scarcely been addressed in previous Cox’s proportional hazards regression model was used to
studies.9,11 assess the effects of patient and caregiver related factors on
We wished to examine the relationship between survival survival. P values less than 0.05 were considered as
(with diagnosis, age and gender taken into consideration) on significant.
the one hand and clinical features and method of receiving
HMV on the other. We also asked if the size of the treating
center and the HMV treatment prevalence in the county did Results
impact survival as a possible evidence of differences in
caregiver experience. We hypothesized that larger center The majority (75%) of the patients was ventilated during
size and/or higher treatment prevalence would be asso- night-time only, 22% were ventilated 8–24 h/day and a small
ciated with longer survival. proportion (3%) was ventilated in day-time only. Ventilatory
support was in all cases delivered by intermittent positive
pressure ventilators. From therapy start HMV was provided
Methods as non-invasive ventilation (NIV) via a mask in 91% and as
tracheostomy invasive ventilation (TIV) in 6% (some missing
Our database was the Swedish HMV Register.12,13 Inclusion data). Still, in the patients starting HMV electively 2% only
criteria: all Swedish patients prescribed long-term ventila- were tracheostomized while this was the case in 19% of the
tion for domiciliary use. Patients using only CPAP for any acute patients. Information on transition from NIV to TIV (or
purpose or ventilators for physiotherapy only were not the opposite) does not with certainty reach the register.
included. For patients starting therapy after January 1, 1996 Demographic information and the distribution of patient/
we prospectively register data on inter alia blood gases, caregiver related factors in the diagnostic groups are
concomitant oxygen therapy, ventilator interface, mode and displayed in Table 1. Although the Pickwick group is by far
daily duration of therapy, and acute vs. elective initiation of the biggest of the eight diagnostic groups the neurological
HMV. The register continuously obtains data on vital status patients including ALS patients make a total of almost the
and the dates of death from the Swedish Population same quantity as the Pickwick patients.
Register. The distribution of the patient-related factors with
In this study we analyzed only patients over the age of 18 respect to the two caregiver related factors are shown in
years when starting HMV ðN ¼ 1526Þ during the 10-year Table 2. It can be seen that acute start of HMV and the use of
period 1996–2005. The primary outcome in this study was concomitant oxygen were more frequent in small centers
death (before March 15, 2006). and in the counties with low HMV prevalence. Counties with
In the preliminary analysis of survival we divided the low prescription rates of HMV also had a higher proportion of
patients according to the underlying diagnoses into eight tracheostomies.
groups: 1. Pickwickian (in the register defined as sleep The survival of the entire material is shown in a
apnea syndrome with respiratory insufficiency), 2. pulmon- Kaplan–Meier plot in Fig. 1. The ALS patients are evidently
ary disease (66% COPD patients), 3. neurological disease different from all other groups by having the poorest
other than amyotrophic lateral sclerosis (ALS), i.e. mainly survival rate at all times; that is a probability to survive
patients with progressive neuromuscular disease (88%), 4. after 2 years of 20% and after 5 years of just above 5%. The
ALS, 5. post-poliomyelitis syndrome, 6. scoliosis, 7. tuber- relative risk for death in the ALS patients compared to all
culosis sequelae (TBC) and finally 8. other diseases, a group other patients was 8.02 (CI 6.48–9.92, Po0:001).
consisting of patients with high cervical lesions, central Among the remaining groups the pattern is more uniform;
hypoventilation and a medley of other diagnoses. In view of still, mutual differences are present e.g. the TBC patients’
the pronounced difference in survival between the ALS and relative risk for death compared to the scoliosis patients was
non-ALS patients and the small differences within the latter 1.91 (95% CI: 1.25–2.90, Po0:01).
group, we joined the non-ALS patients as one group for The relationship between survival and the singled out
further analyses. patient and caregiver related factors when starting HMV is
Among the nearly 50 clinics that manage the treatment of shown in Tables 3 and 4. Table 3 displays the univariable
HMV patients, we arbitrarily label as ‘‘big centers’’ those six analyses for each factor in the ALS and non-ALS groups. None
clinics who reported more than 100 patients starting HMV of the factors significantly affected survival in the ALS
during the 10 years inclusion period. These centers cared for patients, we, therefore, proceeded to multivariable analysis
55% of the HMV patients. For at least 8 years from register in the non-ALS patients only (Table 4).

Please cite this article as: Laub M, Midgren B. Survival of patients on home mechanical ventilation: A nationwide prospective study.
Respiratory Medicine (2006), doi:10.1016/j.rmed.2006.10.007
 ARTICLE IN PRESS
Survival in HMV 3

Table 1 Primary data in 1526 adults starting HMV.

Pickwick Pulmonary Neuromusc ALS Post-polio Scoliosis TBC Other

N (% of all patients) 422 (28) 251 (16) 224 (15) 165 (11) 141 (9) 123 (8) 98 (6) 102 (7)
Start age (SD) 61 (11.2) 63 (11.5) 49 (16.2) 64 (10.5) 67 (9.1) 62 (13.2) 73(6.0) 56 (16.2)
Age475 years (%) 7 11 3 13 18 18 41 12
Male gender (%) 56 37 58 68 48 36 41 58
PaCO2 (SD) 7.1 (1.3) 7.7 (1.6) 7.3 (1.7) 6.6 (1.4) 7.1 (1.3) 7.5 (1.3) 7.6 (1.0) 7.2 (1.6)
PaO2 (SD) 7.5 (1.5) 7.4 (1.8) 9.0 (2.0) 9.7 (1.7) 8.3 (1.4) 7.7 (1.8) 7.4 (1.3) 8.0 (2.2)
PaO2o7.4 kPa (%) 49 53 22 8 22 42 49 37
Acute start (%) 31 33 32 17 19 27 19 39
TIV (%) 1 4 19 4 8 3 2 21
Big center (%) 53 54 57 70 57 44 61 44
Top county (%) 33 31 25 10 12 24 43 19

Arterial blood gases were obtained in elective patients only. Big center (%) is the fraction of the patients cared for by big centers. Top
county (%) is the fraction of the patients living in one of the four high treatment prevalence counties.
TIV: tracheostomy invasive ventilation.

Table 2 Percentage distribution of patient-related Table 3 Univariable analyses of relative risk for death
factors at HMV initiation with respect to the two in the ALS and non-ALS patients.
caregiver related factors.
ALS Non-ALS
Big center Top county
Age475 years
Yes No Yes No 1.30 2.05***
0.77–2.21 1.61–2.62
Age475 years 12 12 14 12 Male gender
Male gender 54 48 53 51 0.79 0.92
Concomitant O2 15 31*** 15 25*** 0.54–1,16 0.76–1.10
PaO2o7.4 kPa 40 37 43 37 Concomitant O2
Acute start 22 37*** 24 30* 2.0 1.85***
TIV 6 8 3 8** 0.89–4.61 1.52–2.25
PaO2o7.4 kPa
TIV: tracheostomy invasive ventilation. 0.78 1.27*
*Po0.05. 0.31–1.98 1.04–1.56
**Po0.01.
Acute start
***Po0.001.
1.29 1.76***
0.83–2.03 1.45–2.13
TIV
0.70 1.51**
0.28–1.71 1.12–2.03
Big center
1.31 0.86
0.88–1.95 0.72–1.03
Top county
1.02 1.13
0.59–1.75 0.93–1.37

Each cell shows on top the hazard ratio and below the 95%
confidence intervals.
TIV: tracheostomy invasive ventilation.
*Po0.05.
**Po0.01.
***Po0.001.

As seen age was a significant factor for survival while

gender was not. Among the patient related factors we found
that concomitant oxygen therapy, TIV and acute start of HMV
were independently associated with worse survival. Con-
Figure 1 Probability to survive in 1526 adult patients after cerning the two caregiver-related factors we found no
starting HMV. relationship to survival.

Please cite this article as: Laub M, Midgren B. Survival of patients on home mechanical ventilation: A nationwide prospective study.
Respiratory Medicine (2006), doi:10.1016/j.rmed.2006.10.007
 ARTICLE IN PRESS
4 M. Laub, B. Midgren

very close to the findings in our study. In a control group

Table 4 Multivariable analyses of relative risk for death
consisting of 52 patients who renounced ventilatory support,
in the non-ALS patients.
none were alive after 2 years. In a recent study3 one group
Age475 years 2.05*** of ALS patients showed a survival rate similar to our findings,
1.55–2.70 while survival was better in another group that underwent
Male gender 1.06 early systematic respiratory evaluation (a caregiver related
0.86–1.31 factor). In our study we were unable to demonstrate any
Concomitant O2 1.63*** factors affecting survival in the ALS patients, besides the
1.27–2.10 diagnosis itself. This indicates that the rapid loss of motor
PaO2o7.4 kPa 1.04 neuron function in ALS overshadows the relatively strong
0.83–1.30 prognostic factors in the non-ALS group.
Acute start 1.79*** Hypoxia as well as need of concomitant oxygen have
1.19–2.70 previously been depicted as factors that were related to
TIV 1.79** poor survival in HMV.9,17 We could not confirm that hypoxia is
1.19–2.68 a significant factor probably because hypoxia in many cases
Big centre 1.03 is corrected by HMV.18 On the other hand, we could confirm
0.83–1.27 that the use of concomitant oxygen was associated with
Top county 1.22 worse prognosis indicating that these two factors are not
0.97–1.52 equivalent. The use of concomitant oxygen may reflect a
component of pulmonary parenchymal disease (with worse
Each cell shows on top the hazard ratio and below the 95% prognosis). However, the interpretation that it reflects
confidence intervals. suboptimal ventilator therapy remains a possibility. Smaller
TIV: tracheostomy invasive ventilation. centers with less experience may be more prone to correct
*Po0.05. hypoxemia with oxygen than to adjust the ventilator
**Po0.01.
settings to achieve the desired therapy goal. Furthermore,
***Po0.001.
launch of HMV in an acute clinical setting involved a greater
risk for death, a fact Duiverman et al.10 found in a
miscellaneous group of 20 mainly TBC patients as well.
Discussion An explanation of why acute start of HMV, use of
concomitant oxygen and TIV were more frequent in small
From a large material of adult patients on HMV recorded centers and in the counties with low HMV prevalence may be
prospectively in a national register we have in this report that clinical procedures and followup prior to HMV start
presented 10-year survival data. We have shown that were less extensive compared to that in big centers or top
patients suffering from ALS at all times distinctly had the counties. This may result in more patients embarking on
poorest probability to survive. Among the remaining patient ventilatory support in an emergent situation and in turn
groups differences in survival could be pointed out, the best entailing more patients starting this as TIV. Need of
long-term survival rates being twice as good as the poorest, concomitant oxygen at HMV start may be related to this
but the overall picture was relatively homogeneous. We scenario or may reflect suboptimal ventilator therapy as
have, therefore, with respect to impact factors on survival discussed above.
analyzed them as one group. Although this may be an Most studies of to day report outcome of HMV delivered as
oversimplification, it enables us to perform these analyses in NIV. In this study, 98 patients (6%) from start were embarked
a much larger and more comprehensive material. We looked upon TIV and this was coupled with a greater risk for death,
at a number of patient related factors when HMV was which may not be surprising as one might claim that the
launched and in the non-ALS patients we found that use of disease severity was more profound in these patients. On
concomitant oxygen, TIV and start in an acute clinical the other hand, as we discussed above in the ALS patients,
setting were factors which were associated with a greater tracheostomy per se should not necessarily imply shorter
hazard for death. Although there were differences in clinical survival and in some patients TIV is more feasible compared
practice between small and big centers and between high to NIV. Post polio and scoliosis patients on TIV have,
and average prevalence counties, these differences seemed furthermore, been reported to perceive better health
to have no effects on survival. compared to NIV for these diagnoses.19 It is, however, a
In ALS previous studies have stated varying survival rates. weakness of our study that the register contain no certain
In an early series of 101 patients14 survival rates were 69% at documentation on transition from NIV to TIV, yet our
2 years and 33% at 5 years. These are considerably higher assumption is that this was performed in less than 10% of
rates compared with more recent smaller studies and our all patients and most often in the neurological patients.
report. Part of an explanation of this may be that all One could expect that treatment in big centers with a
patients were on TIV, whereas this only applies to 5% of our much greater patient flow or in centers with more consistent
ALS patients (at therapy start). TIV may be preferable in followup procedures resulted in a better outcome, but in our
patients with severe paralyses, especially of bulbar nature. report this was not the case. One study has shown a possible
We also speculate that caregivers and relatives may be more connection to consistent procedures9 as the survival was
hesitant to withdraw TIV than NIV.15 In another study16 better in the patients (primarily on LTOT) with complete
including 122 patients, 38 used NIV more than 4 h/day and data of respiratory function. On the other hand, if the
had a survival rate after 2 years just below 20%, which is patient selection in the centers is biased according to local

Please cite this article as: Laub M, Midgren B. Survival of patients on home mechanical ventilation: A nationwide prospective study.
Respiratory Medicine (2006), doi:10.1016/j.rmed.2006.10.007
 ARTICLE IN PRESS
Survival in HMV 5

expertise and tradition, most of the centers, small or big, 5. Gustafson T, Franklin KA, Midgren B, Pehrsson K, Ranstam J,
might have worked up a somewhat uniform quality of care Ström K. Survival of patients with kyphoscoliosis receiving
for the patients once HMV has been started. We have, mechanical ventilation or oxygen at home. Chest, in press.
furthermore, previously shown13 that with regard to elective 6. Hill NS. Noninvasive ventilation for chronic obstructive pulmon-
patients the counties with high treatment prevalence had as ary disease. Respir Care 2004;49(1):72–87.
strict indications to launch HMV as all other counties, 7. Leger P, Bedicam JM, Cornette A. Nasal intermittent positive
pressure ventilation. Long-term follow-up in patients with
demonstrating that disease severity at HMV launch is quite
severe chronic respiratory insufficiency. Chest 1994;105:100–5.
homogeneous in Sweden. As regards center size and 8. Simonds AK, Elliott MW. Outcome of domiciliary nasal inter-
treatment prevalence our hypothesis was that ‘‘big is mittent positive pressure ventilation in restrictive and obstruc-
better’’, a notion we can turn down and recast to ‘‘size tive disorders. Thorax 1995;50:604–9.
doesn’t matter’’. 9. Chailleux E, Fauroux B, Binet F, Dautzenberg B, Polu JM.
In conclusion, in a large material of adult HMV patients, Predictors of survival inpatients receiving domociliary oxygen
the ALS group had by far the poorest survival rate. Survival therapy or mechanical ventilation. A 10-year analysis of
was roughly in line with that in previous studies and ANTADIR observatory. Chest 1996;109:741–9.
predictors of increased risk for death in non-ALS patients 10. Duiverman ML, Bladder G, Meinesz AF, Wijkstra PJ. Home
mechanical ventilatory support in patients with restrictive
were greater age, use of supplemental oxygen treatment,
ventilatory disorders: a 48-year experience. Respir Med
tracheostomy ventilation and launch of HMV in an emergent
2006;100:56–65.
clinical setting. The relative risk for death was not affected 11. Calvert LD, McKeever TM, Kinnear WJM, Britton JR. Trends in
by the size of the treating center and the treatment survival from muscular dystrophy in England and Wales and
prevalence in the county. impact on respiratory services. Respir Med 2006;100:1058–63.
12. Midgren B, Olofson J, Harlid R, Dellborg C, Jacobsen E,
Nørregaard O. Home mechanical ventilation in Sweden, with
Acknowledgments
reference to Danish experiences. Respir Med 2000;94:135–8.
13. Laub M, Berg S, Midgren B for the Swedish Society of Chest
The Swedish HMV Register is supervised by the Swedish Medicine. Home mechanical ventilation in Sweden—inequali-
Society of Chest Medicine and is acknowledged as a quality ties within a homogenous health care system. Respir Med
register by the Swedish health authorities. The register has 2004;98:38–42.
been given financial support by the Swedish Board for Health 14. Oppenheimer EA. Amyotrophic lateral sclerosis: care, survival
and Welfare and from BREAS Medical AB. There are no other and quality of life on home mechanical ventilation. In: Robert
financial interests between the industry and the authors. D, Make BJ, Leger P, et al., editors. Home mechanical
ventilation. Paris: Arnette Blackwell; 1995. p. 249–60.
15. Borasio GD, Voltz R. Discontinuation of mechanical ventilation
References in patients with amyotrophic lateral sclerosis. J Neurol
1998;245:717–22.
1. Annane D, Chevrolet JC, Chevret S, Raphaël JC. Nocturnal 16. Kleopa KA, Sherman M, Neal B, Romano GJ, Heiman-Patterson T.
mechanical ventilation for chronic hypoventilation in patients Bipap improves survival and rate of pulmonary function decline
with neuromuscular and chest wall disorders. The Cochrane in patients with ALS. J Neurol Sci 1999;164:82–8.
Database of Systematic Reviews 2000;(1):Art. No.: CD001941. 17. Muir JF, Girault C, Cardinaud JP, Polu JM. Survival and long-term
DOI:10.1002/14651858.CD001941. follow up of tracheostomized patients with COPD treated home
2. Simonds AK. Home ventilation. Eur Respir J 2003;22(Suppl. mechanical ventilation: a multicenter French study in 259
47):38s–46s. patients. French Cooperative Study Group. Chest 1994;106:
3. Farrero E, Prats E, Povedano M, Martinez-Matos JA, Manresa F, 201–9.
Escarrabill J. Survival in amyotrophic lateral sclerosis with 18. Laub M, Midgren B. The effects of nocturnal home mechanical
home mechanical ventilation. The impact of systematic ventilation on daytime blood gas disturbances. Clin Physiol
respiratory assessment and bulbar involvement. Chest 2005; Funct Imaging 2006;26:79–82.
127:2132–8. 19. Markström A, Sundell K, Lysdahl M, Andersson G, Schedin U,
4. Jeppesen J, Green A, Steffensen BF, Rahbek J. The Duchenne Klang B. Quality-of-life evaluation of patients with neuromus-
muscular dystrophy population in Denmark, 1977–2001: pre- cular and skeletal diseases treated with noninvasive and
valence, incidence and survival in relation to the introduction invasive home mechanical ventilation. Chest 2003;122:
of ventilator use. Neuromuscul Disord 2003;13:804–12. 1695–700.

Please cite this article as: Laub M, Midgren B. Survival of patients on home mechanical ventilation: A nationwide prospective study.
Respiratory Medicine (2006), doi:10.1016/j.rmed.2006.10.007