A prospective longitudinal multicentre study of health related quality of life in ICU survivors with COPD

Survival after intensive care of patients with an acute exacerbation of chronic obstructive pulmonary disease (AECOPD) is poor. Almost 50% of patients die within two years of admission [1‐3]. Important considerations for patients, their families and healthcare providers are not only survival but also health-related quality of life (HRQL) after intensive care [4]. There are only a few published reports on HRQL after intensive care in this group of patients [2, 5]. In the COPD and Asthma Outcome Study (CAOS), HRQL was assessed in 291 AECOPD patients 6 months after ICU care in the United Kingdom using the EuroQoL 5-Dimension (EQ-5D) and 20-item Airways Questionnaire (AQ20) instruments [5]. These patients' HRQL scores were significantly lower than those in an age-matched general population, but similar to those of outpatients with COPD. Rivéra-Fernández and coworkers followed HRQL for six years in fifty patients with AECOPD at Spanish ICUs using a self-developed validated instrument [6]. They found a nonsignificant decline in HRQL scores in this selected cohort. These results from the United Kingdom and Spain were surprising to us, given the facts that COPD is a progressive disease and HRQL in COPD deteriorates in tandem with increasing disease severity and age [7]. Hence, the present study was undertaken to examine the development of HRQL from six months up to twenty-four months after intensive care. We hypothesized that HRQL declines with time in COPD patients after intensive care. We used the EQ-5D instrument and the Short Form 36 Health Survey (SF-36), both of which have been validated and are commonly used in evaluating the critically ill patient population [8‐11]. The results were contrasted using measurements of HRQL in two reference groups, one with COPD and another without COPD, both derived from the general population in the same region as the studied ICUs.

In this prospective, longitudinal, multicentre study, we recruited patients between August 2000 and July 2004 at three mixed medical-surgical ICUs in southeastern Sweden. One is a university hospital and two are general hospitals. Each ICU admits 500 to 750 patients annually. Nearly all admissions to these three ICUs are emergencies, and the most common primary diagnoses are multiple trauma, sepsis, and disturbances in the respiratory or circulatory system or both. The Linköping University Regional Ethics Review Board approved the study protocol.

Patients ages 18 years and older who were admitted to and remained in the ICU for more than 24 hours (N = 1,663) were sent a letter inviting them to participate in a follow-up of their health. A set of structured questionnaires was mailed to survivors, who had given their consent to participate in the study, at six, twelve and twenty-four months after discharge from the ICU. Replies were initially received from 980 patients (59%). Their mean age was 58.2 years, and 42% were female. Selected results from this follow-up (LIVA Health-related quality of life after intensive care in adults Study) have been published previously [12].

We identified patients with COPD from among patients with respiratory disease in the LIVA Study (COPD-ICU group) by matching their data with information in the Swedish Hospital Discharge Register and the Cause of Death Register using the unique Swedish personal identity number. Patients with primary codes from the Tenth Revision of the International Classification of Diseases (ICD-10) that were compatible with COPD (J44.0, J44.1, J44.8 or J44.9) from hospital admission up to five years before study start were included in the COPD-ICU group. Patients who died within five years of the study start and had a primary or secondary cause of death compatible with COPD (ICD-10 codes cited above) on their death certificates were also included in the COPD-ICU group. Thus, 51 patients with COPD were identified using the criteria described above. Their HRQL scores were followed for 24 months after discharge from the ICU. Data for loss of responders are shown in the study flowchart in Figure 1. The clinical databases in each hospital were used to extract data on age, sex, reason for admission, Acute Physiology and Chronic Health Evaluation II (APACHE II) score, ICU length of stay (LOS) and time in hospital, time spent on ventilation and survival. We used the ICD-10 codes of prior hospital admissions found in the Swedish Hospital Discharge Register to compute the Charlson comorbidity index (CCI) score [13].

To contrast HRQL measurements in the COPD-ICU group, we used data from a public health survey performed in the same geographic region. The survey was mailed to 10,000 randomly selected individuals ages 20 to 74 years, from among whom 6,093 (61%) responded [14]. We formed a COPD reference group (N = 38) by matching the personal identity numbers of survey responders with information in the Swedish Hospital Discharge Register and the Cause of Death Register using the same ICD-10 codes as those used for the COPD-ICU group.

The patient characteristics were split into those who were followed throughout the study and those who were lost to follow-up (Table 1). Twenty patients (39%) did not complete the study: nine died and eleven did not respond to questionnaires at 12 or 24 months (see study outline in Figure 1). Survival, which was followed for two years longer than HRQL, was roughly 40% at 48 months (Figure 2).

There were three important findings in this prospective, longitudinal, multicentre study. First, measures of HRQL in COPD patients remained low but stable from six to twenty-four months after discharge from the ICU. Thus, we could not confirm our hypothesis of a decline in HRQL in this cohort of chronically ill patients. Second, HRQL in the COPD-ICU patient cohort was not significantly different from patients with COPD who had been hospitalized but not treated in ICU, that is, the COPD reference group. Third, HRQL among ICU patients who responded at six months but subsequently died was poor.

Awareness of HRQL after critical illness is important if we are to avoid the notion that ICU treatment of COPD patients is futile and a burden for the patient and to society [38, 39]. HRQL is difficult to forecast in critically ill COPD patients who survive ICU care, because information that is typically available, such as age, sex, ICU LOS, severity of illness and lung function parameters, is of limited value for predicting HRQL after recovery [30, 35‐37].

The present study showed that HRQL in COPD survivors after intensive care was lower than that in the general population but quite stable, and that it did not decline from six to twenty-four months after discharge from the ICU. HRQL at 24 months was similar to that of patients with COPD who had not received ICU treatment.