Mortality due to infectious diseases in Spain has been falling over the last few decades, but pneumonia remains the primary cause of death in this group (35.9%).1 The most common type of pneumonia, and the one that causes most hospitalizations, is community-acquired pneumococcal pneumonia (CAPP),2,3 the incidence of which remains high in Spain and the rest of Europe.4

Alcohol use disorders (AUDs) are a well-recognized risk factor for CAPP and have an impact on complications and outcomes of patients during admission.5–13 AUDs increase the risk of developing sepsis during pneumonia, a complication that prolongs hospital stay, worsens prognosis14 and increases the rate of unscheduled readmissions.15

We explored this problem in patients aged 18–74 years, admitted to a selected group of 87 Spanish hospitals between 2008 and 2010, and attempted to control for other confounding and interaction factors, such as age, sex, type of hospital, other addictions and comorbidities. The aim of this study was to analyze the potential influence of AUDs on mortality, prolonged stay and increased costs among patients admitted due to CAPP.

Pneumococcus remains the most common causative agent of community-acquired pneumonia. In the United States, however, the rate has declined, and is now only detected in 10%–15% of hospitalized cases. This reduction has been attributed to the combined influence of various factors,22 such as widespread administration of the pneumococcal polysaccharide vaccine in adults,23 almost universal administration of the pneumococcal conjugated vaccine in children,24 and falling smoking rates.25 In Spain and other European countries where these vaccines have been used less and where high smoking rates persist, pneumococcus is still responsible for the greatest proportions of community-acquired pneumonia.4,26

The results obtained in this study indicate that AUDs have a considerable impact on in-hospital mortality in CAPP patients, causing significantly prolonged hospital stays and generating additional costs per discharged patient. Both occasional excessive consumption of alcohol and chronic alcohol abuse or dependency syndromes cause serious disturbances in specific and non-specific immunity, generating a risk factor not only for CAPP but also for the serious complications observed in these patients.11,14,27

Due to size of the sample and the diversity of the hospitals in this study, our results are generalizable and need not be limited to patients from one or a few hospitals. This, to the best of our knowledge, is the first study conducted in Spain that calculates excess mortality, prolonged stay and additional costs attributable to AUDs in CAPP patients.

The major challenge in analyzing the influence of AUDs on the prognosis and other outcomes of hospitalized patients is adequate control of the confounding bias. Length of stay, costs, and in-hospital mortality differ depending on the reason for admission, the severity of the disease, comorbidities, type of hospital, and other demographic and social patient characteristics.28 Including the hospital group in the multivariate model for controlling the confounding bias is very important, since scientific evidence has shown that quality of medical care and outcomes differ depending on hospital type, facilities and standards of care.1,6,7,10

Our study has several limitations. Our data were sourced exclusively from the MBDs, and no additional patient data were provided. Throughout the study, the definitions of addictive disorders, CAPP and comorbidities were used exactly as assigned by the doctors in each hospital, and coded and entered into the databases by the data managers, with no knowledge of interhospital variability. ICD-9 codes for identifying CAPP are those internationally used for studies based on hospital discharge databases, but there is no means of corroborating these diagnoses with any clinical, radiological and laboratory criteria for diagnosing CAPP that may be mentioned in the clinical records. Previous studies have found that ICD-9 code 481 is highly sensitive and highly specific for hospitalized cases of CAPP if all and not only the principal diagnostic codes are included. This is because CAPP patients are frequently admitted with sepsis, respiratory failure or other diagnoses that are coded as the principal diagnosis, and the diagnosis of CAPP is relegated to one of the secondary diagnoses.17,18,29 To avoid this data bias, all diagnostic codes were taken into consideration in this study, not only the principal diagnosis. Moreover, patients with programmed admission were excluded to reduce as far as possible the inclusion of cases with possibly nosocomial pneumococcal pneumonia.

Databases such as MBDs also have notable advantages. The data collected is usually entered in nearly all hospital discharge records. As all cases are included, quite accurate estimates can be made on incidence, prevalence, comorbidities, complications, and mortality of the diseases seen.9,30,31 These data can be analyzed retrospectively, unlike other designs that need prospective data collection. Data can be collected relatively quickly and easily over long periods and from a large number of patients, as was the case in our study, and such systematic collection considerably reduces costs. In these studies, less selection bias may be generated than in studies in which patients or their legal representatives may refuse to provide written informed consent to participate. Another considerable advantage lies in the availability of defined costs for each DRG stratified by hospital group and year, which facilitates the calculation of excess costs due to CAPP and AUDs.

A consensus document produced by several Spanish scientific societies recommends anti-pneumococcal vaccination in adults with underlying disease, including patients with AUDs.32 This recommendation should be following in all care settings, including hospital departments in which patients with AUDs are identified.

The results of this study are a reminder that the diagnosis of alcohol, tobacco and drug abuse and the introduction of therapy should be one of the main therapeutic objectives prior to discharge of CAPP patients. Investigating the alcohol and drug use and smoking habit of each patient is the ethical and professional duty of each physician. Strategies such as a brief talk on the risks of alcohol, smoking and drugs, and a note in the discharge report advising the primary care physician of the problem in case the patient needs to be referred to specialized detoxification units, have been shown to be effective33–36 in preventing complications and readmissions. Reducing the number of admissions and readmissions attributable to these disorders would help to reduce the costs of sick leave from work and hospital stays, increase the availability of beds in hospitals and reduce the risk of death. Each case of CAPP associated with alcohol, smoking or drug addiction disorders that is avoided also reduces the overall burden borne by these patients and their families.

In patients admitted for CAPP, AUDs account for an excess in-hospital mortality of 50.8%, prolongation of stay of 2.3 days, and excess costs of €1869.20.

This study was funded by: (a) the Government Delegation for the National Drug Plan. Ministry of Health, Social Services and Equality (Grant No. 2009I017, Project G41825811), and (b) Subsidies for Funding Biomedical and Health Science Research in Andalusia for 2013. Department of Health and Social Affairs of the Government of Andalusia (PI-0271-2013).

All authors are equally contributed to the study design, data analysis and interpretation, drafting and review of the article and approval of final version.

The authors state that they had no conflict of interests.