Background
Community-acquired pneumonia (CAP) is a common and potentially serious illness acquired outside a hospital or long-term care facility or other recent contact with the health care system [
1].
Improving the care of adult patients with CAP has been the focus of many different organisations, and several have developed guidelines for the treatment of CAP. In North America, CAP guidelines were first developed in 1993, and in 2007, the joint Infectious Diseases Society of America (IDSA) / American Thoracic Society (ATS) guidelines [
2,
3] were proposed. The process, which began in the United States and Canada, has now been implemented in numerous countries throughout the world [
4]. In the Gulf region, the IDSA/ATS guidelines are considered as one of the most important practice guidelines for the diagnosis and management of CAP. Previous studies have shown that implementing IDSA/ATS guidelines for hospitalised CAP patients in general wards is associated with good patient outcomes [
5‐
7]. For instance, a study conducted in the United States has evaluated the impact of use of the IDSA/ATS therapeutic guidelines among 35 477 inpatients with CAP. This study showed that, after adjustment for severity of illness and other potential confounders, guideline-concordant therapy was associated with a decrease of in-hospital mortality (odds ratio [OR], 0.70; 95 % confidence interval [CI], 0.63–0.77) and of sepsis (OR, 0.83; 95 % CI, 0.72–0.96), with a reduction in both length of stay and duration of parenteral therapy by approximately 0.6 days (
p < 0.001 for both comparisons) [
5]. Several studies on the outcomes of patients with CAP and adherence to the 2007 IDSA/ATS guidelines [
2] have been published [
8‐
10]. However, to our knowledge, only one published study in a Gulf Arab state (Oman) [
11] has been conducted to compare the inpatient management of CAP with established clinical guidelines (the Gulf Cooperation Council CAP guidelines (GCC CAP) [
12]).
The main objectives of our study were to describe treatment practices in CAP patients admitted to hospitals (but not to intensive care units) in five countries in the Gulf region (United Arab Emirates, Kuwait, Bahrain, Oman and Qatar) and to compare the findings to the 2007 IDSA/ATS treatment guidelines.
Discussion
This study was a multicentre, multinational, observational study which aimed to describe treatment practices in CAP patients admitted to hospitals in five countries in the Gulf region (United Arab Emirates, Kuwait, Bahrain, Oman and Qatar) and to compare the findings to the 2007 IDSA/ATS treatment guidelines. To our knowledge, this is one of the first studies to assess treatment standards, which has included a large sample of patients to be conducted in the Gulf region.
Overall, our study shows that the management of CAP in Gulf Arab countries participating in the study is in line with the IDSA/ATS guidelines. For example, information on cough and fever as one of the clinical criteria for CAP was collected systematically and chest radiography showed new infiltrative changes in the majority of patients (70.6 %).
However, for the rest of the population, a radiogram was not performed or did not show new infiltrative changes.
The type, route of administration and duration of antimicrobial therapy was also generally consistent with the IDSA/ATS guidelines for hospital ward treatment. Antimicrobial therapy used in the Gulf region was in general consistent with the recommendation of starting initial therapy empirically with fluoroquinolones until laboratory results could be obtained to guide more specific therapy. All patients received therapy initiated immediately on the day of hospitalisation, with fluoroquinolones being the most widely used treatment. However, ten percent of the study population received a combination of a fluoroquinolone and a β-lactam during hospital admission, which is not the treatment recommended by the IDSA/ATS expert panel for patients not admitted to an ICU [
2].
The pathogens responsible were not isolated and characterised in the majority of patients and thus represent the major divergence between current practice in the Gulf region and the IDSA/ATS guidelines. Characterisation of CAP pathogens was only achieved in 12 % of patients, principally by analysis of sputum samples. In addition, information about use of Gram’s staining method or any other methods used to identify pathogens, was not collected in our study. The most frequently identified pathogen was
Streptococcus pneumoniae, which is consistent with findings from other countries, as noted in the IDSA/ATS guidelines [
2].
In 95 % of cases, the initial empiric therapy in our study was not switched, suggesting that this empiric therapy was effective. However, since the responsible pathogen was not isolated in the majority of the patients, it is not known whether the treatment was the most appropriate in these patients. Once the patient was clinically stable, around 80 % of patients were switched to an oral antibiotic within three days (72 h) as recommended in the guidelines.
A further objective was to evaluate treatment outcome at hospital discharge and up to approximately 3 weeks after discharge. At hospital discharge, criteria for treatment success were fulfilled by the majority of patients. However, around one quarter of patients were discharged before their leukocyte count had normalised. In addition, it should be noted that at hospital discharge, complete radiographic recovery was only detected in around one-third of patients. Clinical signs related to CAP were resolved in 97.9 % (n = 183) at the last follow-up visit, performed around 3 weeks after discharge.
Our study identifies four areas in which current practice for management of CAP in the Gulf region falls short of international guidelines and could be improved.
The first corresponds to isolation and characterisation of the pathogenic agent, which needs to be performed systematically. The lack of pathogen isolation in the majority of patients observed in the study may result in inappropriate antimicrobial therapy being prescribed and thus expose patients to unnecessary risk.
The second relates to the failure to perform chest radiography systematically at the post-discharge follow-up visits, with around 90 % of patients not being assessed. This may result in residual infection in the lungs being missed and increase the risk of relapse if the physician or patient decides to interrupt antimicrobial treatment. Nonetheless, in everyday clinical practice in North American and Europe, it is also common for follow-up chest radiographs to be performed only if symptoms persist or there is a clinical suspicion of infection [
16].
The third relates to the fact that only one quarter of patients attended a second follow-up visit 3 weeks after discharge. It would be important to encourage physicians to collect information from their patients, either during a formal hospital visit, and particularly if residual clinical or radiological signs of infection were detected at the previous follow-up visit, or by telephone, 3 weeks after discharge to ensure that the patient is effectively cured.
Finally, 10 % of the study population received a combination of a fluoroquinolone and a β-lactam during hospital admission, which is the treatment recommended by the IDSA/ATS expert panel for patients directly admitted to an ICU [
2] and not for patients included in our study (
ie patients not directly admitted to ICU). Consequently, in those patients the choice of CAP antimicrobial therapy was not consistent with the guidelines. Neverthless, at hospital admission, seventeen percent of patients (
N = 117) were classified as moderate to high risk for mortality according to the Fine criteria [
13] and consequently could be considered by the physician to be eligible for an ICU regimen without necessarily being directed to one. Because the overall efficacy of many classes of antimicrobial regimens is considered to be good, over-treatment use is expected to produce the same clinical outcome but with an increased risk of antiobiotic resistance and side effects. For this reason, the selection of antimicrobial regimens for empirical therapy should be based on prediction of the most likely pathogen and knowledge of local susceptibility patterns with a preference given to the more potent drugs because of their benefit in drecreasing the risk of selection for antiobiotic resistance [
2].
We compared treatment practices in CAP patients in the Gulf region to the IDSA/ATS guidelines since they are the most used guidelines worldwide, are well known and used by doctors of the region and are one of the recommended standard of care in the Gulf region.
The adherence to regional guidelines (
ie GCC CAP guidelines [
12]) has already been evaluated in Oman [
11]. This analysis showed that there was very poor adherence to the guidelines with respect to CAP severity assessment and provision of preventive measures upon hospital discharge and highlighted the necessity to improve the implementation of such guidelines in the region. Several studies have evaluated the level of adherence to the IDSA/ATS guidelines. For instance, a recent study from Italy performed in 191 patients with CAP revealed that the adherence to antibiotic treatment guidelines was poor, since only 47 % of patients received an empirical antibiotic regimen that was adherent to guidelines [
17].
The enrolled patients were principally at low mortality risk (82.9 %) according to the Pneumonia Severety Index (PSI) developed by Fine [
13], which assigns points based on age, laboratory parameters, and the presence of different comorbidities. The PSI is one of the instruments recommended by the IDSA/ATS panel in order to assess mortality risk and to identify patients with CAP who may be candidates for outpatient treatment [
2]. It has been suggested that patients with risk class I and II should be treated as outpatients, risk class III patients should be treated in an observation unit or with a short hospitalisation, and risk class IV and V patients should be treated as inpatients [
13]. In the present analysis, we decided not to include patients with risk class I. We assumed that those patients did not require a particular hospital monitoring and are most often candidates for outpatient treatment than that of other classes.
The proportion of patients considered at low mortality risk (PSI II and III) was higher than that anticipated and also higher than the rate reported in the literature. For example, it has been reported that the rate of low-risk patients hospitalised due to CAP in Israel, USA and Western Europe was around 30 %[
13,
14,
18,
19]. The most likely explanation for this is that patients at high risk, who in fact died during hospitalisation, are not captured in our study.
A principal limitation of our study is the lack of information on methods used for pathogen identification and the failure to characterise the pathogen in the majority of patients, which compromises the quality of the data collected. In particular, this may have led to less frequent, but epidemiologically important, pathogens not being detected.
A second major limitation relates to the mode of recruitment of the patients, which precludes collecting information on in-hospital mortality. We were thus unable to document mortality rates in our cohort, in particular in patients with moderate to high risk mortality according to the Fine criteria [
13]. According to the literature, the rate of mortality has been estimated to be low in the Gulf region compared to other areas of the world. For instance, the mortality rate among patients with pneumoniae has been estimated at 3.5 % (
N = 6/172 patients) in Oman [
11] compared to 7.2 % (
N = 147/ 2039 patients) in Europe [
20].
A third limitation relates to the low mortality risk of the majority of the patients enrolled. This may be a consequence of exclusion of patients who required hospitalisation in an ICU and failure to capture high risk patients who died in hospital. In the sample size calculation, it was anticipated that 50 % of patients would be at high risk, whereas in fact >80 % of patients were at low risk. This enrichment in low-risk patients may have led to an over-optimistic view of treatment outcome in clinical practice, which may not apply to more severely-ill patients. In addition, it is possible that standards of care, in terms of adherence with IDSA/ATS guidelines, may have been inferior in patients who died.
Finally, although we examined data from five Gulf countries represented by 38 hospitals, the majority of the study population (65 %;
n = 444) came from the United Arab Emirates, which may introduce some bias. To a certain extent, the number of patients included reflects the demographic weight of each country, with UAE having the highest population (8 million) in 2010, followed by Kuwait and Oman (3 million each), Qatar (1.8 million) and Bahrain (1.2 million) [
21]. However, the UAE is somewhat over-represented and Qatar and Oman somewhat under-represented. In part this may be explained by the number of hospitals available for recruitment. The federal structure of the UAE has led to the development of multiple autonomous health services in each of the constituant emirates leading to a larger number of hospitals
per capita than in the other Gulf countries. In contrast, in Oman and Qatar, the population, and health care provision, is concentrated in the capital city. The number of patients recruited in Oman was particularly low (28). In comparison, a study performed only in Oman [
11] in a single centre and over a similar study period to that of our study (2 years), included a total of 170 patients. It is possible that there was competition for patient recruitment between the two studies. Moreover, all participating countries have large populations of migrant workers whose socioeconomic status differs from the local population. Socioeconomic status may influence morbidity risk and adherence with treatment following discharge and this was not taken into account in our analysis. Finally, it should be noted that our study did not include Saudi Arabia, which is the largest country of the Gulf region. Consequently, our results should not be generalised to the entire Gulf region.
Competing interests
BM, AAZ, OMAA, RA and MSN have received honoraria from Sanofi for their contributions to the G-TinCAP study. REB is an employee of Sanofi, which funded the study. The authors declare that they have no competing interests.
Author contributions
All authors developed the hypotheses and conceived the analysis plan. All authors contributed to the analysis and interpretation of the study results. All authors participated in the revision of the manuscript and agree to be accountable for all aspects of the work. All authors approved the submission of this final draft. All authors read and approved the final manuscript.