Effectiveness of Influenza Vaccination for Individuals with Chronic Obstructive Pulmonary Disease (COPD) in Low- and Middle-Income Countries

Abstract

Chronic obstructive pulmonary disease (COPD) is one of the leading causes of death globally. In addition to the mortality associated with it, people with COPD experience significant morbidity, making this set of conditions a major public health concern. Infections caused by influenza virus are a preventable cause of morbidity and vaccination has been shown to be effective. The evidence of their benefit in persons with COPD mainly comes from high-income countries where influenza vaccination is used in routine practice, but little is known about the effectiveness, cost-effectiveness, and scalability of vaccination in low- and middle-income countries. We therefore systematically reviewed and present evidence related to vaccination against influenza in persons with COPD with a special focus on studies from low- and middle-income countries (LMICs). Available data from 19 studies suggest that the use of influenza vaccine in persons with COPD is beneficial, cost-effective, and may be relevant for low- and middle-income countries. Wider implementation of this intervention needs to take into account the health care delivery systems of LMICs and use of prevalent viral strains in vaccines to be most cost effective.

Keywords:

• Chronic Obstructive Pulmonary Disease (COPD)
• influenza vaccine
• effectiveness
• low and middle income countries

Background

Chronic Obstructive pulmonary disease (COPD) is among the top 10 leading causes of death, worldwide (1). It is a large public health concern not just due to the mortality associated with it but also substantial morbidity. COPD accounts for 4.6% of all disability adjusted life years (DALYS) lost globally in 2010 (1). COPD is seen commonly among smokers and is characterised by an irreversible, progressive airflow obstruction. Some patients with moderate-to-severe COPD are prone to frequent exacerbations (three or more exacerbations per year) that are an important cause of hospital admission and readmissions, affecting health care costs for COPD (2). Acute exacerbations are particularly common in persons older than 60 years of age (3).

Acute exacerbations may be bacterial or viral in aetiology. Rhinovirus is most commonly associated with these exacerbations, though, in severe exacerbations that require admission to hospital, influenza virus is most common (4). Vaccines to prevent these acute exacerbations due to influenza may therefore be important to decrease the excess mortality and morbidity associated with COPD. The influenza vaccine has been found to provide moderate protection against virologically confirmed influenza, although this protection is greatly reduced or absent in some seasons. The evidence is consistent for efficacy in young children (aged 6 months to 7 years) and in adults aged 65 years or older (5). There is a paucity of studies that evaluate the effectiveness of influenza vaccines in COPD alone, and furthermore, these are mainly from high-income countries. In this review, we present the evidence for effectiveness of influenza vaccines in persons with COPD with a special focus on low and middle income countries.

Methods

We searched the PubMed, EMBASE, and the Cochrane databases for studies evaluating the effectiveness of influenza vaccination in adults with COPD. The search terms used were [“Influenza Vaccine” OR “flu vaccine”] AND [“COPD” OR “chronic obstructive lung disease”]. We included randomized controlled trials (RCT), non-randomized controlled trials, single-arm pre-post evaluations, and observational studies (cohort and case-control) for evaluation of effectiveness. We also included studies about vaccination coverage and determinants of vaccination among persons with COPD and have discussed this separately. We also hand-searched reference lists from review articles to try to capture a complete list of included studies. We excluded studies in which people with COPD were less than 10% of the target population.

Studies that were published on or before June 20, 2014 and in English language were included. Data from eligible studies was extracted, categorized, and narratively synthesized. The World Bank classification of countries (high, middle, or low-income) from 2014 was used to categorize countries' in which each study originated.

Vaccines commonly used against influenza and included in this review

Most of the vaccines currently being used are inactivated trivalent (IIV3) and have been in use since they were first introduced in 1978 (6). IIV3 contains a mixture of three influenza viruses: one influenza A (H1N1) virus, one influenza A (H3N2) virus, and one influenza B virus. Each year, the World Health Organization (WHO) recommends the composition of influenza vaccines, based on the results of global influenza surveillance and the prediction of the strains most likely to circulate in the forthcoming influenza season (7). The inactivated influenza vaccine (IIV3) and live-attenuated influenza vaccine (LAIV) are currently most widely in use. There are three types of IIV: whole virus vaccines, split virus vaccines, and subunit vaccines. IIVs are typically split virus or subunit vaccines, and are produced from highly purified, egg grown influenza viruses (8). The LAIV is administered via the intranasal route, and the LAIV strains replicate in the epithelial cells of the nasopharynx. LAIV induce strain-specific IgA production, and these IgA responses are associated with protection against influenza illness (9). Other recent advances in achieving a greater and more sustained immune response include the use of adjuvants, use of cell culture lines, high-dose vaccines, intradermal vaccines (ID) and the recent inactivated quadrivalent vaccine (IIV4) (8).

Results

The search yielded 189 studies after removing duplicates from each database. There were 28 articles that met eligibility for inclusion. Among these, 19 were reports of intervention trials, case control or cohort studies and 9 were cross-sectional that studied vaccination coverage and its determinants. Among the studies included (Table 1), 9 were randomized controlled trials and 10 analytic studies. Most of the studies were conducted in high-income countries, namely: Korea (1), Japan (4), Spain (1), USA (5), Scotland (1), Poland (1); and the following upper and lower middle income countries: India (1), Thailand (3), Turkey (1), and Taiwan (1).

Table 1. Studies included for review

	Authors	Year	Country	Study design (Sample size)	Outcome measure
1	Seo et al. (21)	2014	Korea	Case control (828)	Hospitalization
2	Montserrat-Capdevila et al. (22)	2014	Spain	Retrospective cohort (1323)
3	Kiyohara et al. (23)	2013	Japan	Cohort – secondary data January 1995–December 2009 (202 338 COPD deaths)	Mortality
4	Anar et al. (15)	2010	Turkey	Cross-sectional (82)	Antibody titres, seroconversion rates and seroprotection and Hospitalization
5	Chuaychoo et al. (11)	2010	Bangkok	Randomized Trial comparing vaccines (160)	Antibody titres, seroconversion rates and seroprotection
6	Schembri et al. (24)	2009	Tayside, Scotland	Cohort -Secondary data from The Health Improvement Network (THIN) database (77,120)	Mortality
7	Ayabe et al. (25)	2009	Japan	Cohort (289)	Pneumonia and acute exacerbations of CLD
8	Sumitani et al. (26)	2009	Japan	Retrospective cohort additive inoculation of influenza vaccine (IIV) and 23-valent pneumococcal vaccine (PV)	Hospitalization
9	Furumoto et al. (27)	2009	Japan	Trial comparing vaccine types (167)	Pneumonia and acute exacerbations of CLD
10	Menon et al. (14)	2008	India	Cohort (87 males)	Number of ARI classified as outpatient treatment, hospitalizations and need for mechanical ventilation
11	Wang et al. (16)	2007	Taiwan	Retrospective cohort (35,637)	Mortality
12	Karwat et al. (28)	2007	Poland	Cohort (178 asthma and COPD)	Pneumonia and acute exacerbations of CLD
13	Kositanont et al. (13)	2004	Thailand	RCT (123)	Laboratory diagnosed influenza caused illness (LDI)
14	Wongsurakiat et al. (12)	2004	Bangkok, Thailand	RCT comparing vaccine and placebo (125)	Number of ARI classified as outpatient treatment, hospitalizations and need for mechanical ventilation
15	Gorse et al. (29)	2003	US VA hospital	RCT comparing vaccine types (2215)	Laboratory diagnosed influenza caused illness (LDI)
16	Gorse et al. (30)	1997	US, VA hospital	RCT comparing vaccine types (29)	Pulmonary function status and clinical status
17	MRC advisory group (31)	1977	US	3 trials (111) RCT comparing vaccine and placebo	Antibody titres, seroconversion rates and seroprotection
18	Bartsch et al. (32)	1977	US	Live influenza vaccines RIT 4025 and RIT 4050, containing the recombinants of A/Scotland/840/74 and A/Victoria/3/75 with A/PR/8/34 virus	Pulmonary function status and clinical status
19	Prevost et al. (33)	1975	US	2 trials (44) Of different virus strains administered intranasally (LAIV)	Antibody titres, seroconversion rates and seroprotection
20	Fell et al. (34)	1977	US	Double blind RCT	Self-assessments of cough, breathlessness and wheeze
21	Howells and Tyler (35)	1961	US	RCT	Pneumonia and acute exacerbations of CLD and Pulmonary function status and clinical status

CLD- Chronic lung disease, ARI- Acute respiratory illness, RIT4025, 4050- vaccine strains.

The 9 randomized trials compared vaccine with placebo or vaccines with different routes of administration and a placebo. There was a large variation in the outcome measures used for determining the effectiveness of vaccine. Most of the studies used vaccines that were matched to the circulating strain in the season the study was conducted as per the WHO recommendations for that year. Only two studies (21, 22) report findings during an epidemic of influenza, rest were all seasonal reports. These studies included persons with a diagnosis of COPD, most often based on the GOLD (Global Initiative for Chronic Obstructive Lung Disease) criteria using Forced Expiratory Volume by spirometery and did not exclude any stage of the disease. The findings of these studies are summarised categorized by study design, as randomized trials (Table 2) and as observational studies (Table 3).

Table 2. Outcomes and findings from randomized trials

Sample no	Study	n	Comparison	Outcome	Findings	Conclusion
Serological Outcome Measures and Pulmonary Function Tests (PFT)
1	Chuayachoo et al. (2010) (11)	160	Intradermal and Intramuscular vaccines	Geometric mean titres (H1N1)	Prevaccination (ID, IM) 23.5,199.1 Post vaccination (ID, IM) 22.1, 223.2	≥4-fold increase at 4 weeks Both routes effective
2	MRC advisory group (1977)	111	Live Vaccine and saline placebo	Actual serial titres reported	50% showed 4-fold increase in titres in vaccinated and none in unvaccinated	Effective increase in titres
3	Bartsch et al. (1977) (11)	-	2 live attenuated vaccines of different strains	Pulmonary function tests (PFT) 4 weeks	No change in PFT	Both vaccines were ineffective in improving PFT
4	Prevost et al. (1975) (33)	40	2 different strains of Intranasal vaccines	Seroconversion rates	86 and 73% for each vaccine in patients with low titres pre vaccination	Both had satisfactory antibody response
5	Gorse et al. (1997) (30)	29	Bivalent live attenuated influenza A virus vaccine (CAIIV) or saline solution placebo along with i.m. injection of IIV in both	Pulmonary function tests and antibody titres	No change in PFT Mean levels of anti-HA immunoglobulin A (IgA) antibodies in nasal washings increased significantly after vaccination with CAIIV and IIV compared to placebo	Increase in antibody titres but no change in PFT and clinical status
Clinical Outcome Measures
1	Kositanont et al. (2004) (13)	123	Live attenuated Vaccine and placebo	Lab diagnosed influenza	8.2% in vaccine group 27.4% in placebo	Significantly lower influenza in vaccinated
2	Wongsurakiat et al. (2004) (12)	125	Split virus vaccine and placebo	Influenza related ARI	6.8 and 28.1 per 100 persons years in vaccine and placebo groups	Significantly lower ARI in vaccinated
3	Gorse et al. (2003) (29)	2215	CAIVV- vaccine co administered with IIV3 and with placebo	Laboratory diagnosed Influenza Efficacy as1 minus hazard ratio	4.5% and 5.3% in vaccine and placebo groups No difference in Kaplan Meir curves, similar efficacy of 0.16 (95% CI: 0.22, 0.43)	Vaccine not compared to placebo but both vaccines effective
4	Furumoto et al. (2009)	167	23-valent PV + IIV and only IIV	Pneumonia and acute exacerbation of Chronic Lung Diseases (CLD)	0.53 episodes per patient year of infectious exacerbations 10.3% in PV+IIV and 26% in IV	Significantly lower exacerbations in the PV+IIV group compared with the IIV group but no difference for pneumonia

ARI-Acute Respiratory Infections, CAIVV-Cold-Adapted Influenza Virus Vaccine, IIV3- Trivalent Inactivated Influenza Virus Vaccine, PV- Pneumococcal Vaccine, IIV – Inactivated Influenza Vaccine.

Table 3. Outcomes and findings from observational studies

		Study		Design			N	Outcome measure	Finding	Conclusion
Hospitalizations and Acute Exacerbations Aas Outcome Measures
1		Seo et al. (2014) (21)	Case control, hospital controls matched 1:1 for age, gender, and date of hospital visit.			828		Hospitalization	Lower hospitalization among vaccinated: 33.7% (95% [CI] 14.0–49.0%; P = 0.002)	Significant reduction in hospitalizations
2		Montserrat-Capdevila et al. (2014) (22)	Retrospective cohort			1323		Hospitalization	Lower hospitalization among vaccinated: 68.4% (95%CI: 47.5–81.0)	Significant reduction in hospitalizations
3		Anar et al. (2010) (15)	Cross-sectional			82		Acute exacerbations requiring hospitalizations Antibody response	38.6% in vaccinated and 76.3% in unvaccinated, 50% of vaccinated had Increase IgG and IgA	Significantly fewer hospitalizations in vaccinated
4		Ayabe et al. (2009) (25)	Cohort			289		Acute exacerbation wheezing, pneumonia	11 of 189 (5.8%) in the vaccinated group and 23 of 100 (23%) in the unvaccinated group (RRR: relative risk reduction 74.7%, 95% CI: 0.5–0.87).	Significant reduction in acute exacerbations
5		Sumitani et al. (2009) (26)	Retrospective cohort additive inoculation of influenza vaccine (IIV) and 23-valent pneumococcal vaccine (PV)			-		Bacterial respiratory infections, number of hospitalizations, and length of hospital stay in the 2 years prior to and after PV inoculation.	3.16 vs. 1.95 infections; (p=0.0004) and 0.79 vs. 0.43 hospitalizations; (p=0.001) favour of vaccine	Significant reductions in acute infections
6		Menon et al. (2008) (14)	Cohort			87		Episodes and severity of ARI	The incidence of ARI 28.6 per 100 person-years prior to vaccination and 9.7 per 100 person-years post vaccination [relative risk (RR) 0.33; p = 0.005	Significant reductions in ARI
7		Karwat et al. (2007) (28) Karwat et al. is 2007 in Table 2 but 2006 in reference. Please correct year.</AQ>	Cohort			178		Acute exacerbations	The number of exacerbations decreased more than twice comparing the time before and after immunization: 1.7 ± 1.5 and 3.9 ± 2.4 per year, respectively	Reduction in acute exacerbations
Mortality as Outcome Measure
1	Kiyohara et al. (2013) (23)		Cohort – secondary data before and after law for vaccination		202 338 COPD deaths			Mortality from COPD	in January (RR 0.84; 95% CI 0.81-0.88), February (RR 0.85; 95% CI 0.81-0.89) and March (RR 0.92; 95% CI 0.88-0.96) among the population aged ≥ 65 years.	Reductions in mortality in elderly attributed to vaccination. However, in the population aged <65 years, statistically no significant changes in the COPD mortality rate were found in any month after the amendments were made
2	Schembri et al. (2009) (24)		Secondary data from The Health Improvement Network (THIN) database		177,120 patients with COPD			All-cause Mortality relative risks	Vaccinated – 0.30 (95% CI 0.28 to 0.32) and 0.98 (95% CI 0.96 to 1.11) and Unvaccinated - 0.59 (95% CI 0.57 to 0.61) during the influenza season and 0.97 (95% CI 0.94 to 1.00) outside the season	Significant reduction in all-cause mortality among vaccinated compared to not vaccinated
3	Wang et al. (2007) (16)		Retrospective cohort		35,637 patients with COPD			Mortality due to COPD in vaccinated	HR = 0.55	45% reduction in mortality in those vaccinated

Randomized trials

There were five studies that reported serological measures and clinical status as outcomes. These studies compared different strains of live attenuated vaccines, intranasal vaccine, different routes of administration (intramuscular and intra dermal). All studies report up to a 4-fold increase in antibody response. Only 2 studies, Bartsch et al. (32) and Gorse et al. studying live attenuated vaccines of different strains reported no improvement in clinical status and pulmonary function tests.

There were four randomized trials which assessed clinical outcomes such as lab diagnosed influenza, influenza related Acute Respiratory Illness (ARI) and acute exacerbation of chronic lung disease. All four studies reported a significantly lower rate of these outcomes in the vaccinated group.

Observational studies

There were seven studies that reported hospitalizations and acute exacerbations as outcomes. All these studies reported significantly reduced outcomes in the vaccinated group when compared with those not vaccinated. There were three studies that reported mortality data and all these three, too, reported significant relative reductions in mortality rates ranging from 45% to 50%.

Discussion

Findings from both trials and observational studies consistently show benefit from vaccination in this high-risk group of persons with COPD, although the level evidence is not very strong due to the heterogeneity of studies and low number of patients included. This is in concurrence with the recommendation from a Cochrane review, which evaluated the role of influenza vaccine in COPD patients and included only randomized trials (10). Across 7 studies reporting on hospitalizations, vaccination was associated with 33.7% to 74.7% relative reductions in incidence of hospitalizations and the 3 studies reporting mortality as an outcome showed 18% to 45% relative reductions among those vaccinated.

The evidence from low and middle income countries is sparse and there are no trials conducted in these countries to guide policy. The 6 studies from low and middle income countries have findings consistent with the evidence from high-income countries. The three studies from Thailand report a 4-fold increase in geometric mean titres of antibodies after vaccination (11) and 19.2% to 21.3% lower influenza related illness in the vaccinated (12,13). The studies from India and Turkey report a 37.7% to 76% relative reduction in severity of episodes (14) and hospitalizations (15), respectively. The only study from low and middle income countries that assessed mortality as the outcome, from Taiwan, reported a 45% relative reduction in those vaccinated (16).

A recent review and meta-analysis assessed the role of influenza vaccines in low- and middle-income countries, specifically for vulnerable populations. They concluded that in COPD patients, there are statistically significant protective effects from using inactivated influenza vaccine in terms of reduced laboratory-confirmed influenza (efficacy 70% (95%CI: 24–88)), ILI (effectiveness 66% (95%CI:12–87)), and COPD-related hospitalization (effectiveness 71% (95%CI: 17–90)), but not COPD-related outpatient visits (effectiveness 50% (95%CI: −60 to 84)) (17).

Cost effectiveness

There is only one study from Thailand (18) that has conducted an economic evaluation of influenza vaccine compared to placebo (12) specifically in persons with COPD. The study reports the incremental cost effectiveness ratio of vaccination from a health provider perspective by applying the direct medical costs to the results obtained in the RCT. More than 90% of the costs of treatment of influenza-related ARI were costs of hospitalization and for patients with moderate and severe airflow obstruction; more than 90% of these costs were attributed to mechanical ventilation.

The study reports that for every 100 patients with mild COPD, the cost would be 24,840 baht (approx. 720 US dollars) more and would prevent 18.2 outpatients, 4.8 hospitalizations, and 0 mechanical ventilations due to ARI related to influenza. Also, for every 100 moderate COPD and every 100 severe COPD patients, vaccination would have prevented 5 outpatient visits, 6 hospitalizations, and 6 moderate cases needing mechanical ventilation and 21 outpatient visits, 4 hospitalizations, 8 severe cases requiring mechanical ventilations, respectively. Vaccination was considered highly cost effective in prevention of influenza-related acute respiratory illnesses irrespective of severity, and more so in those with moderate to severe COPD (18).

A recent review (2013) of economic evaluations for influenza vaccines summarised data from low- and middle-income countries (19). They included 9 studies, of which 5 were economic modelling studies and all were from upper middle income countries (Thailand, Colombia, China, Brazil, Argentina and Mexico). Overall, the articles included indicated that seasonal influenza vaccination is cost-effective in elderly, children with high-risk conditions, infants, health care personnel and COPD patients in these middle income countries. These findings are similar to evidence from high-income countries (19).

Scalability and feasibility in LMICs

Despite the evidence and the fact that vaccination for influenza in persons with COPD is part of several recommendations in high-income countries, the coverage remains dismally low. There are several patient and provider barriers that have been identified such as no awareness, disbelief in effectiveness of the vaccine, and poor follow-up in general practice. The annual vaccination coverage reported from high-income countries ranged from 30.5% to 82%. (Table 4). The only vaccination coverage study from a middle income country (Turkey) reported that, in 2006, coverage was 14.9%. There is a paucity of studies reporting coverage of influenza vaccination from low- and middle-income countries and hence a comparison with high-income countries was not possible. This limits our understanding of factors that may affect coverage in low- and middle-income countries. Further, implementing vaccination in LMICs in persons with COPD will also be a challenge considering the varied health care delivery systems and its constraints

Table 4. Vaccination coverage and its determinants

Study no	Authors	Year	Country	Study design	Sample	Findings
1	Vandenbos et al.	2013	France	Cross-sectional	476 patients, 246 with COPD	53% coverage
Refusal and intolerance reasons for no vaccination
2	Mowls et al.	2013	Australia	Cross- sectional	3108	31% (confidence interval 1.1, 1.6) more likely to have received an influenza vaccination in the past 12 months compared to those without a breathing test.
3	Ciblak et al.	2013	Turkey	Cross- sectional literature review	27 million at risk groups	14.9% coverage
Disbelief in effectiveness of vaccine
4	Santos-Sancho et al.	2013	Spain	Cross-sectional	European Health Survey (Spain) in 2009/10, data on 15,355 and 1,309 had COPD	49.4% coverage (95% CI: 46.3–52.5%) and 21.3% (95% CI: 20.7–21.9) among people without (p < 0.001).
Factors associated with an increase in vaccination coverage were: being male, perceiving one's health as fair or poor, not smoking, and having seen a doctor during the previous month.
5	Chiatti et al.	2013	Italy	Cross-sectional	5,935	30.5% coverage of adults (N = 670) and 74.8% (N = 2,796) of older people
Age increases the odds of vaccine uptake. Single marital status, Higher-education among the older people, smoking and not having contact with GPs in both age groups, are factors associated with non-vaccination.
6	Vozoris et al.	2009	Canada	Cross-sectional	134,072	47.9% coverage
Men, non-Ontarians, younger age groups, current smokers, and those without a family doctor were less likely to be vaccinated. Vaccine not necessary as reason
7	Schoefer et al.	2007	Germany	Cross-sectional	2131	46.5% coverage
Lack of information, patients' age and education level determinants
8	Jiménez-García et al.	2006	Catalonia, Spain	Cross-sectional	1783	62.5% coverage
Higher age, more frequent contact with the general practitioner and greater length of disease progression considerably increased
9	Jiménez-García et al.	2005	Spain	Cross-sectional	10711	Coverage 87.2%
More frequent contact with the general practitioner and a history of pneumococcal vaccination increase the likelihood of being vaccinated

Also of concern is the matching of prevalent viral strains to the strains used in the manufactured vaccine every year. The studies included in this review do report that the vaccine evaluated was matched to the circulating strain although the specific strains of the season were not always described. The greater effectiveness reported may have been influenced by the dominance of a particular strain. The strategy of vaccination in persons with COPD is more effective and cost effective when the vaccine and circulating viral strains are well matched.

The cost-effectiveness of influenza vaccination depends on the following factors: 1) efficacy and effectiveness of the vaccine, 2) incidence of influenza, 3) vaccination costs, and 4) healthcare costs from influenza (20). Wongsurakiat et al. (12) and others also make the point that the effectiveness of the vaccine in reducing exacerbations will depend on how much influenza-related ARI is present, i.e., whether there is an epidemic or not. These issues will influence the scalability and feasibility of implementing a recommendation for vaccination against influenza among persons with COPD in LMICs. We did not find any studies that assess the costs of innovations in the health system to ensure vaccination in this high-risk population such as tracking and reminding patients which may be required to ensure good coverage.

Conclusion

The available data suggest that the use of influenza vaccine in persons with COPD may be beneficial, cost effective, and relevant for low and middle countries. The implementation will need to take into account the health care delivery systems of LMICs and use of prevalent viral strains in vaccines to be most cost effective.

Funding

We would like to acknowledge the Fogarty International Centre and the Eunice Kennedy Shriver National Institute Of Child Health & Human Development at the National Institutes of Health for the grant number 1 D43 HD065249 which supports the primary author. We would also like to acknowledge the WHO that funded this review (WHO Reference 2013/369613-0).

Declaration of Interest Statement

There are no financial or other relationships of the authors that could have influenced the conduct of this review. The authors alone are responsible for the content and writing of the paper.