Abstract

Many national guidelines recommend annual influenza vaccination of immunocompromised patients, although the decision to vaccinate is usually at clinical discretion. We conducted a systematic review and meta-analyses to assess the evidence for influenza vaccination in this group, and we report our results by etiology. Meta-analyses showed significantly lower odds of influenza-like illness after vaccination in patients with human immunodeficiency virus (HIV) infection, patients with cancer, and transplant recipients and of laboratory-confirmed influenza in HIV-positive patients, compared with patients receiving placebo or no vaccination. Pooled odds of seroconversion and seroprotection were typically lower in HIV-positive patients, patients with cancer, and transplant recipients, compared with immunocompetent controls. Vaccination was generally well tolerated, with variation in mild adverse events between etiological groups. Limited evidence of a transient increase in viremia and a decrease in the percentage of CD4+ cells in HIV-positive patients was found although not accompanied by worsening of clinical symptoms. Clinical judgment remains important when discussing the benefits and safety profile with immunocompromised patients.

Vaccination is an effective means of preventing influenza, and the rapid emergence of oseltamivir-resistant human A(H1N1) seasonal influenza viruses in 2007, which persisted in circulation until mid-2009, highlight the strategic importance of preventative measures [1, 2]. Patients with impaired immune function due to disease or therapy are recommended for annual vaccination in most national vaccination guidelines. These recognize the difficulty in defining a level of immunosuppression that represents a threshold below which vaccination is recommended and above which (all other things being equal) it is not, such as the daily dose of prednisolone producing clinically relevant immunosuppression [3]. Furthermore, the underlying causes of immunocompromise are diverse, and susceptibility to influenza may vary. Decisions about vaccination are frequently devolved in practice to the discretion of individual clinicians, but doubts remain about effectiveness according to etiology of immunocompromise.

We conducted a systematic review and meta-analyses to assess influenza vaccination for immunocompromised patients. Overall results from a public health policy perspective, including details of the study methodology, have been reported elsewhere [4]. This article presents an analysis stratified by etiology of immunocompromise, of potentially greater relevance to clinicians managing such patients.

METHODS

Our full methodology has been detailed elsewhere [4], and an abbreviated protocol is published on the National Institute for Health Research international prospective register of systematic reviews (PROSPERO) Web site [5]. Four review questions were specified a priori, to study the prevention of clinically diagnosed influenza or influenza-like illness (ILI) and laboratory-confirmed influenza, serological response, and adverse events associated with influenza vaccination.

Search Strategy and Selection Criteria

We searched healthcare databases and sources of gray literature for studies reporting outcomes after influenza vaccination in patients with immunocompromise due to primary or secondary immunodeficiency. The definition of these conditions was based on immunization policy of the World Health Organization and the Department of Health, United Kingdom. All identified studies were sifted against the protocol eligibility criteria. Patients with malnutrition and tuberculosis were also included as conditions commonly associated with immunocompromise in developing countries; studies were excluded that reported data from patients with drug-induced immunosuppression where <80% were receiving therapy (including long-term inhaled corticosteroids) [5]. We used a standardized, piloted form to extract data and assessed the risk of bias for all included studies using standard tools [6–8]. Data on serological response and adverse events were extracted according to Committee for Human Medicinal Products (CHMP) criteria, in addition to serious adverse events and disease progression [9, 10].

Data Analysis

This article reports an analysis stratified by etiology of immunocompromise. A narrative approach was used to synthesize the risk of bias assessments and extracted data to inform each review question [11]. Where appropriate, meta-analysis of odds ratios estimated the pooled effect size of vaccinating immunocompromised patients versus immunocompetent controls and immunocompromised patients receiving placebo or no vaccination (PNV) using Stata software, version 10 (StatCorp). Analyses were initially performed using a random effects model, but re-executed using a fixed effects model where heterogeneity was low (I2 < 40%). Meta-analyses were abandoned where heterogeneity was high (I2 > 85%). We assumed statistical significance of pooled odds ratios at the 5% level, as assessed using the χ2 test. Publication bias was assessed using Begg's funnel plots and Egger's regression test.

RESULTS

We identified 209 studies from 219 records meeting protocol eligibility criteria. Details of the study selection process (including PRISMA diagram), a summary of the included studies and risk of bias assessments is described elsewhere [4]. Figure 1 depicts the number of studies identified by etiology of immunocompromise and research question; in addition to the categories of immunocompromise shown, we identified another 7 studies that met the protocol eligibility criteria and were classified as reporting data on “other” etiological categories. However, no subanalysis was undertaken for these owing to heterogeneity between studies and the significant paucity of information. Because of the volume of data generated by this systematic review, forest plots of each meta-analysis have not been presented but are available on request. Meta-analyses of serological outcome measures are summarized in Table 1, and selected results presented in Figures 2 and 3. Publication bias was absent from all meta-analyses unless stated otherwise.

Table 1.

Meta-analysis Results of Serological Outcome Measures by Etiology of Immunocompromise

Outcome Measure Influenza Subtype Comparator Studies Included, No. Pooled ES (95% CI) P for ES I2 (%) P for I2 
Patients with HIV infection        
 SC1 A(H1N1) (S) VICT 17a 0.51 (.40–.66) <.001 19.5 NS 
 SC1 A(H3N2) VICT 15 0.47 (.27–.79) .005 69.4 <.001 
 SC1 VICT 15a 0.34 (.26–.44) <.001 36.7 NS 
 SC1 A(H1N1) (S) PNV 1.62 (.41–6.41) NS 31.2 NS 
 SC1 A(H3N2) PNV 3.99 (.24–65.97) NS 73.1 .05 
 SP A(H1N1) (S) VICT 0.28 (.12–.65) .003 69.8 .003 
 SP A(H3N2) VICT 0.28 (.09–.80) .02 69.7 .003 
 SP VICT 0.24 (.09–.60) .002 64.3 .01 
Patients with cancer        
 SC1 A(H1N1) (S) VICT 12a 0.31 (.22–.43) <.001 34.8 NS 
 SC1 A(H3N2) VICT 12a 0.39 (.21–.71) <.001 66.6 .002 
 SC1 VICT 8a 0.37 (.20–.68) .001 46.5 NS 
 SP A(H1N1) (S) VICT 10a 0.30 (.15–.61) .001 64.3 .002 
 SP A(H3N2) VICT 10a 0.30 (.14–.63) .002 62.6 .003 
 SP VICT 9a 0.30 (.14–.67) .003 68.4 .001 
Transplant recipients        
 SC1 A(H1N1) (S) VICT 10 0.76 (.38–1.51) NS 57.8 .01 
 SC1 A(H3N2) VICT 10 0.38 (.23–.62) <.001 42.8 NS 
 SC1 VICT 10 0.48 (.27–.86) .01 55.2 .02 
 SC2 A(H1N1) (S) VICT 0.52 (.21–1.27) NS 0.0 NS 
 SC2 A(H3N2) VICT 0.28 (.07–1.11) NS 76.5 .01 
 SC2 VICT 0.43 (.08–2.43) NS 87.3 <.001 
 SP A(H1N1) (S) VICT 10 0.28 (.16–.47) <.001 7.3 NS 
 SP A(H3N2) VICT 0.29 (.09–.92) .04 78.4 <.001 
 SP VICT 0.36 (.19–.70) .002 59.0 .01 
Patients with autoimmune disease treated with immunosuppressants 
 SC1 A(H1N1) (S) VICT 0.90 (.45–1.80) NS 68.8 .002 
 SC1 A(H3N2) VICT 1.54 (1.03–2.32) .04 32.1 NS 
 SC1 VICT 0.98 (.43–2.24) NS 65.1 .009 
 SC2 A(H3N2) VICT 1.09 (.13–8.89) NS 53.4 NS 
 SC2 VICT 0.34 (.08–1.37) NS 69.4 NS 
 SP A(H1N1) (S) VICT 0.49 (.29–.84) .01 0.0 NS 
 SP A(H3N2) VICT 0.71 (.43–1.17) NS 3.8 NS 
 SP VICT 9a 0.48 (.22–1.05) NS 65.9 .002 
Patients with respiratory conditions treated with immunosuppressants 
 SC1 A(H1N1) (S) VICT 0.96 (.33–2.79) NS 46.3 NS 
 SC1 A(H3N2) VICT 0.93 (.30–2.85) NS 53.3 NS 
 SC1 VICT 0.29 (.14–.59) .001 23.6 NS 
Outcome Measure Influenza Subtype Comparator Studies Included, No. Pooled ES (95% CI) P for ES I2 (%) P for I2 
Patients with HIV infection        
 SC1 A(H1N1) (S) VICT 17a 0.51 (.40–.66) <.001 19.5 NS 
 SC1 A(H3N2) VICT 15 0.47 (.27–.79) .005 69.4 <.001 
 SC1 VICT 15a 0.34 (.26–.44) <.001 36.7 NS 
 SC1 A(H1N1) (S) PNV 1.62 (.41–6.41) NS 31.2 NS 
 SC1 A(H3N2) PNV 3.99 (.24–65.97) NS 73.1 .05 
 SP A(H1N1) (S) VICT 0.28 (.12–.65) .003 69.8 .003 
 SP A(H3N2) VICT 0.28 (.09–.80) .02 69.7 .003 
 SP VICT 0.24 (.09–.60) .002 64.3 .01 
Patients with cancer        
 SC1 A(H1N1) (S) VICT 12a 0.31 (.22–.43) <.001 34.8 NS 
 SC1 A(H3N2) VICT 12a 0.39 (.21–.71) <.001 66.6 .002 
 SC1 VICT 8a 0.37 (.20–.68) .001 46.5 NS 
 SP A(H1N1) (S) VICT 10a 0.30 (.15–.61) .001 64.3 .002 
 SP A(H3N2) VICT 10a 0.30 (.14–.63) .002 62.6 .003 
 SP VICT 9a 0.30 (.14–.67) .003 68.4 .001 
Transplant recipients        
 SC1 A(H1N1) (S) VICT 10 0.76 (.38–1.51) NS 57.8 .01 
 SC1 A(H3N2) VICT 10 0.38 (.23–.62) <.001 42.8 NS 
 SC1 VICT 10 0.48 (.27–.86) .01 55.2 .02 
 SC2 A(H1N1) (S) VICT 0.52 (.21–1.27) NS 0.0 NS 
 SC2 A(H3N2) VICT 0.28 (.07–1.11) NS 76.5 .01 
 SC2 VICT 0.43 (.08–2.43) NS 87.3 <.001 
 SP A(H1N1) (S) VICT 10 0.28 (.16–.47) <.001 7.3 NS 
 SP A(H3N2) VICT 0.29 (.09–.92) .04 78.4 <.001 
 SP VICT 0.36 (.19–.70) .002 59.0 .01 
Patients with autoimmune disease treated with immunosuppressants 
 SC1 A(H1N1) (S) VICT 0.90 (.45–1.80) NS 68.8 .002 
 SC1 A(H3N2) VICT 1.54 (1.03–2.32) .04 32.1 NS 
 SC1 VICT 0.98 (.43–2.24) NS 65.1 .009 
 SC2 A(H3N2) VICT 1.09 (.13–8.89) NS 53.4 NS 
 SC2 VICT 0.34 (.08–1.37) NS 69.4 NS 
 SP A(H1N1) (S) VICT 0.49 (.29–.84) .01 0.0 NS 
 SP A(H3N2) VICT 0.71 (.43–1.17) NS 3.8 NS 
 SP VICT 9a 0.48 (.22–1.05) NS 65.9 .002 
Patients with respiratory conditions treated with immunosuppressants 
 SC1 A(H1N1) (S) VICT 0.96 (.33–2.79) NS 46.3 NS 
 SC1 A(H3N2) VICT 0.93 (.30–2.85) NS 53.3 NS 
 SC1 VICT 0.29 (.14–.59) .001 23.6 NS 

Abbreviations: CI, confidence interval; ES, effect size; HIV, human immunodeficiency virus; ILI, influenza-like illness; NS, not statistically significant; PNV, placebo or no vaccination; SC1, seroconversion defined as ≥4 fold rise after vaccination; SC2, seroconversion defined as <1:40 to ≥1:40 hemagglutination inhibition titer; SP, seroprotection, defined as ≥1:40 hemagglutination inhibition titer after vaccination; VICT, vaccinated immunocompetent controls.

a One study contributed 2 sets of data included in this meta-analysis.

Figure 1.

Studies identified by etiology of immunocompromise and research question. Subtotals do not add up to 209 because 3 studies recruited multiple groups of immunocompromised patients; 7 studies on “other” causes of immunocompromise are not shown. Abbreviations: AE, adverse events; HIV, human immunodeficiency virus; ILI, influenza-like illness; LCI, laboratory confirmed influenza; SR, serological response. 

Figure 1.

Studies identified by etiology of immunocompromise and research question. Subtotals do not add up to 209 because 3 studies recruited multiple groups of immunocompromised patients; 7 studies on “other” causes of immunocompromise are not shown. Abbreviations: AE, adverse events; HIV, human immunodeficiency virus; ILI, influenza-like illness; LCI, laboratory confirmed influenza; SR, serological response. 

Figure 2.

Selected meta-analysis results of influenza-like illness (ILI) and laboratory-confirmed influenza (LCI) by etiology of immunocompromise. Results represent odds of ILI and LCI compared with placebo or no vaccination; error bars show 95% confidence interval around pooled effect size. Abbreviation: HIV, human immunodeficiency virus.

Figure 2.

Selected meta-analysis results of influenza-like illness (ILI) and laboratory-confirmed influenza (LCI) by etiology of immunocompromise. Results represent odds of ILI and LCI compared with placebo or no vaccination; error bars show 95% confidence interval around pooled effect size. Abbreviation: HIV, human immunodeficiency virus.

Figure 3.

Selected meta-analysis results of serological outcome measures by etiology of immunocompromise. A, Odds of seroconversion with ≥4 fold rise in hemagglutination inhibition titer after vaccination, compared with vaccinated immunocompetent controls. B, Odds of seroprotection with ≥1:40 hemagglutination inhibition titer after vaccination, compared with vaccinated immunocompetent controls; error bars show 95% confidence interval around pooled effect size. Abbreviations: HIV, human immunodeficiency virus; Rx, treated with immunosuppressants.

Figure 3.

Selected meta-analysis results of serological outcome measures by etiology of immunocompromise. A, Odds of seroconversion with ≥4 fold rise in hemagglutination inhibition titer after vaccination, compared with vaccinated immunocompetent controls. B, Odds of seroprotection with ≥1:40 hemagglutination inhibition titer after vaccination, compared with vaccinated immunocompetent controls; error bars show 95% confidence interval around pooled effect size. Abbreviations: HIV, human immunodeficiency virus; Rx, treated with immunosuppressants.

Patients with HIV Infection

Thirteen studies reported data on the incidence of ILI after influenza vaccination in patients with human immunodeficiency virus (HIV) infection. Meta-analysis of 2 studies (PNV comparator) showed a statistically significant pooled odds ratio of 0.20 (95% confidence interval [CI], .05–.88; P = .03). Although significant heterogeneity was present (I2 = 84.7%; P = .01), this result agrees with 2 previous meta-analyses that found a reduction in symptomatic or laboratory-confirmed influenza [12–15]. Among studies unsuitable for meta-analysis, we found evidence to suggest that the rate of ILI after vaccination was lower in HIV-positive patients than in PNV controls, and comparable to that in vaccinated immunocompetent (VICT) controls.

Four studies reported data on laboratory-confirmed influenza in HIV-positive patients (including the 2 aforementioned systematic reviews). Meta-analysis of 2 studies (PNV comparator) yielded a pooled effect size of 0.15 favoring vaccination (95% CI, .03–.63; P = .01) with moderate heterogeneity (I2 = 50.4%; P = NS). Narrative synthesis similarly identified a lower rate of laboratory-confirmed influenza in HIV-positive patients than in PNV controls, although there was a paucity of data comparing these rates with those in VICT controls [13, 16].

Data on serological response to influenza vaccination in HIV-positive patients were extractable from 47 studies. Table 1 shows that meta-analysis of seroconversion and seroprotection by influenza strain typically resulted in a significantly reduced response in HIV-positive patients compared with VICT controls. Analyses of seroconversion for seasonal influenza A(H1N1) and B, compared to PNV controls, favored patients who received active vaccine, but neither analysis reached statistical significance. In addition, some studies reported mean geometric increase in hemagglutination inhibition (HI) titer after vaccination in HIV-positive patients, demonstrating an inferior response compared with HIV-negative controls. Many single-arm clinical studies showed an increase in number of HIV-positive patients reaching the CHMP criteria for serological response after vaccination, compared with baseline.

Twenty-four studies reported local or systemic adverse events, typically mild, because influenza vaccination was generally well tolerated. Where these studies included a control group, variation in the rate of adverse events was shown between HIV-positive patients and VICT or PNV controls, although most found no significant difference or similar frequencies between groups. Thirty-six studies measured laboratory or clinical variables describing the effect of vaccination on HIV status in infected patients, such as CD4+ cell count, HIV load, or RNA level at variable lengths of follow-up. Many single-arm clinical studies and controlled clinical trials reported no significant change in such variables when comparing baseline with post-vaccination levels. In contrast, Gunthard et al described a decrease in proviral DNA levels in patients with <400 RNA copies/mL at baseline but found increased HIV RNA levels at 4 weeks after vaccination in patients with a baseline titer of <50 copies/mL [17]. Other studies reported a transient increase in HIV titer after vaccination, which typically returned to baseline levels within 1 month, but sustained elevations in a number of individual vaccinees [18–21]. In addition, Staprans et al showed a statistically significant correlation between plasma viremia and immunological response, and a substantially greater mean fold increase in HIV RNA titer in individuals with high baseline CD4+ cell count [21]. Rosok et al showed a rapid increase in viremia and modest elevation in the frequency of infected cells of patients with a low baseline HIV load; a decrease in viremia and a significant rise in the rate of infected cells was observed in those with a high baseline HIV load [18]. Two controlled clinical trials showed increases in viremia, and a decrease in the percentage of CD4+ cells was shown by Tasker et al [22, 23]. No serious adverse events were reported as being related to administration of influenza vaccination.

Patients With Cancer

Twelve studies reported data on the incidence of ILI after vaccination in patients with cancer. Meta-analysis of 2 studies (PNV comparator) showed a statistically significant pooled odds ratio of 0.26 (95% CI, .15–.46; P < .001) with no heterogeneity (I2 = 0.0%; P = NS) [24, 25]. Studies comparing rates between patients with cancer and VICT controls showed comparable rates of ILI [24, 26–28]. One study conducted during the 2009 pandemic reported ILI in 5 of 97 patients with cancer, compared with none of 25 healthy controls, although all 5 patients were confirmed negative for influenza A(H1N1) by reverse-transcriptase polymerase chain reaction, and no test of statistical significance was performed [26]. A number of single-arm studies reported low rates of ILI in vaccinated patients with cancer.

Six studies reported on laboratory-confirmed influenza infections in vaccinated patients with cancer, 3 of which recruited a VICT control group. Data were inadequate for meta-analysis, because only 1 study identified a single swab-positive case of influenza A(H3N2) in a vaccinated patient [29].

Fifty studies reported on immunological response to vaccination in patients with cancer. Meta-analysis of seroconversion and seroprotection by influenza strain shows a significantly reduced response in patients with cancer compared with VICT controls (see Table 1). However, evidence of publication bias was found in the analysis for seasonal influenza A(H1N1) (Egger's test; t = −2.37; 95% CI, −6.37 to −.14; P = .04), B (Egger's test; t = −3.39; 95% CI, −9.08 to −1.73; P = .01) and approached significance for A(H3N2) (Egger's test; t = −2.02; 95% CI, −6.00 to .33; P = .07). Multiple single-arm nonrandomized clinical studies showed an increased proportion of patients with cancer attaining the criteria for seroprotection after vaccination. Two nonrandomized controlled trials recruiting immunocompetent control groups concurred with our meta-analyses [27, 30]. The mean geometric increase in HI titer after vaccination in patients with cancer commonly exceeded the CHMP criteria, although to a lesser extent than observed in healthy controls. Immunological response to influenza A/H1N1/California/07/2009 vaccine antigens was assessed in 2 studies demonstrating statistical equivalence compared with healthy controls [31, 32].

Thirty-one studies described local or systemic adverse events that were typically mild, indicating influenza vaccination was well tolerated by patients with cancer. The rate of adverse events was often comparable between study groups, with many studies citing no statistically significant difference compared with healthy controls or PNV groups. No consistent evidence of disease progression or serious adverse events was identified as being related to the administration of influenza vaccination in 8 studies that reported data on these outcome measures.

Transplant Recipients (Solid Organ and Hematological)

Fifteen studies reported the incidence of ILI after vaccination in transplant recipients. Meta-analysis of 2 studies (PNV comparator) showed a statistically significant pooled odds ratio of 0.27 (95% CI, .11–.66; P = .004) with no heterogeneity (I2 = 0.0%; P = NS). Eight single-arm clinical studies showed very low numbers of vaccinated transplant recipients with a diagnosis of ILI (6 reported no cases during follow-up) and another study showed statistically equivalent rates of ILI in those who seroconverted to each vaccine antigen and those who did not [33]. The rates of ILI in studies vaccinating transplant recipients and immunocompetent controls were highly comparable.

Four studies investigated the prevention of laboratory-confirmed influenza after vaccination in transplant recipients. Only 1 study included a control group, and a single case was virologically confirmed from 3 renal transplant recipients who developed ILI after vaccination [34, 35].

Fifty-five studies reported on serological response to influenza vaccination. Meta-analyses showed pooled effect sizes for seroconversion and seroprotection that were lower in patients than in VICT controls (see Table 1), although these only reached statistical significance for seroconversion to influenza A(H3N2) and B subtypes [4]. Meta-analysis of seroconversion (<1:40 to ≥1:40 HI titer after vaccination) for influenza B was conducted but subsequently abandoned because I2 reached 87.3%. Tabulation of studies reporting CHMP parameters for transplant recipients showed findings consistent with our meta-analyses. A weaker response to influenza vaccination was observed compared with VICT controls, although some studies showed comparable or increased responsiveness for some influenza subtypes [36, 37]. Numerous studies described a mean geometric increase after vaccination in excess of the CHMP criteria.

Twenty-five studies described local or systemic adverse events in transplant recipients; influenza vaccination was generally well tolerated with only minor events reported. The rate of adverse events was comparable to that in healthy controls or placebo vaccinees in studies that recruited a control group. Thirty studies reported outcome measures pertaining to tissue rejection or allograft function in transplant recipients vaccinated against influenza, but there was no consistent evidence of an association with these outcomes, nor of serious adverse events.

Patients With Autoimmune Disease Treated With Immunosuppressants

Five studies reported on the rate of ILI in vaccinated patients with autoimmune diseases but were unsuitable for meta-analysis. Limited evidence suggests influenza vaccination is beneficial, resulting in low rates of ILI, possibly comparable to rates in VICT controls.

One study reported on laboratory-confirmed influenza in vaccinated patients with autoimmune disease. Del Porto et al obtained pharyngeal swab samples from all 3 patients with ILI; 1 sample was influenza positive [38].

Twenty-six studies described serological outcome measures in patients with autoimmune diseases. Table 1 shows pooled effect sizes for seroconversion (≥4 fold rise in HI titer; VICT control group) to seasonal influenza A(H1N1) and B were statistically equivalent although significant heterogeneity was present, however the pooled odds ratio was significantly higher for influenza A(H3N2). Meta-analysis of 7 studies for seroprotection against seasonal influenza A(H1N1) gave significantly lower odds in patients compared with VICT controls. The odds of seroprotection against influenza A(H3N2) and B were equivalent between vaccinated patients and VICT controls, although the estimated pooled effect size intimated a weaker response in patients. Additional data from 13 studies support our meta-analyses, although some studies showed a weaker response in patients with autoimmune disease. The mean geometric mean titer after vaccination was commonly greater than the CHMP criteria of 2.5, although less than in VICT controls.

Local or systemic adverse events were described by 21 studies; they were typically mild and self-limited. Rates of adverse reactions were typically similar between vaccinated patients on immunosuppressive therapy and those not receiving therapy or healthy controls. Fifteen studies reported outcomes pertaining to disease progression or worsening of clinical symptoms, although no consistent evidence demonstrated worsening of autoimmune disease activity after vaccination [39–41]. No serious adverse events were reported as being related to administration of influenza vaccination.

Patients With Respiratory Conditions Treated With Immunosuppressants (Predominately Corticosteroids)

We identified a significant paucity in literature describing clinical protection achieved through vaccination of patients with respiratory conditions who are immunosuppressed by receiving oral or inhaled corticosteroids. Two randomized controlled trials showed no cases of ILI in a cohort of patients with pulmonary tuberculosis followed up for 6 months, or of laboratory-confirmed influenza in children infected with tuberculosis followed up for 11 months [42, 43].

Six studies described the immunological response to influenza vaccination in immunosuppressed patients with respiratory conditions. Meta-analysis of seroconversion in 2 studies, with comparison to VICT controls, showed no significant differences for seasonal influenza A(H1N1) and A(H3N2; see Table 1). The pooled effect size for seroconversion to influenza B was 0.29 (95% CI, .14–.59; P = .001). Data from these studies grouped patients receiving oral and inhaled corticosteroids, so we cannot report this separately [44, 45]. Only 1 other study reported serological outcomes based on CHMP criteria (which could not be pooled for meta-analysis), showing an increased rate of seroconversion in patients receiving medium- or high-dose corticosteroids compared with PNV controls [44].

Six studies reported adverse events in patients with respiratory conditions receiving immunosuppressive medication, showing that influenza vaccination was associated with only minor and transient adverse events. However, Nicholson et al reported that differences between symptom scores in vaccine and placebo recipients reached statistical significance for upper respiratory, systemic, and all symptoms [46]. The authors also reported that patients were no more likely to be hospitalized, consult family practitioners, or take antibiotics or oral steroids after administration of active vaccine, compared with placebo. Three studies described the effect of influenza vaccination on asthma symptoms, and 1 study described its effect on tuberculosis antibody levels; none showed a statistically significant difference compared with controls (after statistical adjustment in 1 study) [42, 46–48]. No serious adverse events were reported in relation to the administration of influenza vaccine.

DISCUSSION

This review is the first to report clinical outcome measures following influenza vaccination after stratification by etiology of immunocompromise. Our meta-analyses suggest that vaccinating HIV-positive patients, patients with cancer, and transplant recipients against influenza confers a statistically significant degree of clinical protection against ILI, with additional narrative data in concordance. A paucity of evidence for autoimmune and respiratory conditions treated with immunosuppressants was identified. The pooled odds of seroconversion and seroprotection were commonly lower and reached statistical significance in HIV-positive patients, patients with cancer, and transplant recipients compared with VICT controls, suggesting a suboptimal serological response. However, our meta-analysis for seroprotection in patients with cancer should be interpreted with caution owing to a risk of publication bias. Meta-analyses of serological data for autoimmune and respiratory conditions treated with immunosuppressants were largely statistically equivalent compared with VICT controls (with some exceptions including a likely artifactual result for seroconversion to influenza A[H3N2]). We report 2 meta-analyses of seroconversion in HIV-positive patients compared with PNV controls that show a greater pooled estimate of effect, although neither reached statistical significance. Influenza vaccination was generally well tolerated, although the frequency and distribution of mild adverse effects may differ within and between etiological groups. We identified limited evidence of a transient increase in viremia and a decrease in the percentage of CD4+ cells in HIV-positive patients after vaccination, although this was not accompanied by worsening of clinical symptoms.

We previously reported several limitations of this study, including risk of bias and confounding due to inclusion of nonrandomized and observational studies, pooling of outcome measure data after vaccination against 2009 pandemic influenza A(H1N1) although only one such study was included in the ILI meta-analysis for transplant recipients (PNV comparator), and clinical and statistical heterogeneity that may challenge the validity of our meta-analyses [4]. Additional caveats pertaining to this secondary analysis warrant discussion. Potential confounding variables specific to our 5 etiological groups may not have been controlled or accounted for, such as duration and dose of highly active antiretroviral therapy in HIV-positive patients or chemotherapy in patients with cancer, differences between patients with solid tumors and hematological malignancies or those currently receiving chemotherapy, stem cell and tissue transplant recipients, and variation between patients receiving different antirejection regimens after transplantation. Pooling data from studies of patients with respiratory conditions treated with inhaled or oral steroids may be problematic and reduces the reliability of our findings because of a potential difference in the long-term immunosuppressant effect associated with different steroid dose equivalents and routes of administration. We note the lack of adequate data to support analyses of response to vaccination in different groups of patients with cancer, and the preponderance of cancer studies included in our review, which recruited children with hematological malignancies. Including studies published since the 1970s is potentially problematic due to advancements in antiretroviral and immunosuppressive therapies, including the increased severity of induced immunosuppression caused by current chemotherapy regimens. Previous exposure to influenza vaccination, timing of vaccine administration and immunosenescence may also be important effect modifiers contributing to heterogeneity. We did not attempt to delineate adverse events that are associated with immunological or other laboratory markers of disease not directly affecting the immunocompromised state but may otherwise precede a clinically significant event (eg, alloreactive antibody levels in transplant recipients). Furthermore, our review questions and search strategy were not designed to resolve potential differences in degree of immunocompromise or in the effect of vaccination within any of the etiological groups we analyzed. Many included studies stratified patients by CD4+ cell count, HIV RNA load, or AIDS status, which prevented an assessment of response to vaccination for all recruited patients. Finally, control subjects recruited in some studies had current or previous risk factors for HIV, raising the possibility of subclinical disease.

In light of our findings, and in recognition that immunocompromised patients are at higher risk of infection or serious complications after infection [3, 49], we advocate that clinicians should follow national guidance and consider prioritizing such patients for influenza vaccination. Being mindful of the caveats above and the breadth of each review question, clinicians should consider the generalizability of our findings to their typical patient population and setting.

In conclusion, clinical judgment remains important when discussing the benefits and safety profile of influenza vaccination with immunocompromised patients, which should include consideration of underlying comorbid conditions and risk factors for exposure to infection or influenza-related disease. Our analyses show evidence of effectiveness and a favorable safety profile for the 5 etiological groups reported. However, our data suggest that the clinical and serological response to influenza vaccination in patients with HIV infection, patients with cancer, and transplant recipients may be worthwhile but inferior compared with VICT controls, whereas patients with autoimmune or respiratory conditions receiving immunosuppressive drug therapy may show a response similar to that in controls. We acknowledge that the degree of protection conferred by vaccination may vary by etiology, and within each etiological group according to immune status or other factors, although our systematic review was not designed to study these. There is a paucity of evidence from patients with autoimmune or respiratory conditions on immunosuppressive drug therapy. Because of the methodological limitations of currently published research, we advocate further well-designed interventional studies to improve the overall quality and quantity of evidence on the clinical and serological response to influenza vaccination in immunocompromised patients [4].

Notes

Acknowledgments. C. R. B. and J. S. N.-V.-T., the primary and senior authors, respectively, take responsibility for the work and act as guarantors of the data. We acknowledge and thank the following for their support and advice throughout the project: Dr Charles Penn and Sara Martins (Global Influenza Programme, World Health Organization [WHO]); Dr Gayle Dolan (Health Protection Agency North East). We thank the European Vaccine Manufacturers, GlaxoSmithKline, Novartis, and Sanofi Pasteur MSD for responding to our request for literature potentially relevant to this systematic review.

Author contributions. Protocol design: C. R. B., B. C. M., R. C. H., J. S. N.-V.-T.; execution of search strategy and screening: C. R. B., B. C. M., A. B. H., R. C. H., University of Nottingham Influenza and the ImmunoCompromised (UNIIC) Study Group; risk of bias assessments and acquisition of data: R. C. H., UNIIC Study Group; Analysis and interpretation of data: C. R. B., A. B. H., J. S. N.-V.-T.; manuscript preparation and contribution of intellectual content: C. R. B., B. C. M., A. B. H., R. C. H., J. S. N.-V.-T.; final manuscript approval: C. R. B., B. C. M., A. B. H., R. C. H., UNIIC Study Group, J. S. N.-V.-T.

UNIIC Study Group membership. University of Bielefeld, Germany, A. Zanuzdana; National Health Service (NHS) Blackpool, United Kingdom, G. Agboado; University of Nottingham, United Kingdom, E. Orton, J. Enstone, R. Puleston, Y. Vinogradova; WHO, Switzerland, L. Béchard-Evans, M. Mukaigawara, J. Peñalver; Health Protection Agency, United Kingdom, G. Morgan, C. Stevenson, R. Weston, G. Dabke, S. Haroon, C. Hird, J. Roberts, L. Lingard; NHS Leicestershire County, United Kingdom, G. Augustine, L. Ahyow; Brighton and Sussex University Hospital NHS Trust, United Kingdom, M. Butt; Harvard University, United States of America, S. Kim, J. Figueroa; NHS Nottingham City, United Kingdom, R. Howard, J. O'Boyle; Wessex Deanery, United Kingdom, M. O'Brien; NHS Derbyshire County, United Kingdom, H. Denness; Mersey Deanery, United Kingdom, S. Farmer; Solihull NHS Primary Care Trust, United Kingdom, P. Fisher; Imperial College London, United Kingdom, F. Greaves; Leicester Partnership Trust, United Kingdom, M. Haroon; Lancaster University, United Kingdom, R. Isba; University College London, United Kingdom, D. A. Ishola, M. Kerac; NHS Northamptonshire, United Kingdom, V. Parish; NHS Halton and St Helens, United Kingdom, J. Rosser; NHS Nottinghamshire County, United Kingdom, S. Theaker, D. Wallace; NHS East Lancashire, United Kingdom, N. Wigglesworth; Tokyo Medical and Dental University, Japan, H. Horiuchi.

Financial disclosure. This study was commissioned by the Global Influenza Programme, WHO. The funder specified the research questions but had no other role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The University of Nottingham Health Protection Research Group (J. S. N.-V.-T., C. R. B., B. C. M., A. B. H., J. Enstone, R. Puleston) is an official WHO Collaborating Centre for pandemic influenza and research. It receives limited funding from WHO in support of specific activities. The sources of funding for each author during this study were: East Midlands NHS Healthcare Workforce Deanery, United Kingdom (C. R. B., B. C. M.); University of Nottingham (A. B. H.); Global Influenza Programme, WHO (R. C. H.); Health Protection Agency, United Kingdom (J. S. N.-V.-T.).

Potential conflicts of interest. The University of Nottingham Health Protection Research Group is currently in receipt of research funds from GlaxoSmithKline. The group has recently accepted an unrestricted educational grant for influenza research from F. Hoffmann–La Roche. Research on influenza funded by an unrestricted educational grant from AstraZeneca is also underway. The aforementioned funding received from GlaxoSmithKline, F. Hoffmann–La Roche, and AstraZeneca did not support any aspect of this study. J. S. N.-V.-T. has received funding to attend influenza-related meetings, lecture and consultancy fees and research funding from several influenza antiviral drug and vaccine manufacturers. All forms of personal remuneration ceased in September 2010, but influenza-related research funding from GlaxoSmithKline, F. Hoffmann–La Roche and AstraZeneca remains current. He is a former employee of SmithKline Beecham (now GlaxoSmithKline), Roche Products, and Aventis-Pasteur MSD (now Sanofi-Pasteur MSD), all before 2005, with no outstanding pecuniary interests by way of shareholdings, share options or accrued pension rights. Among UNIIC Study Group members, A. Zanuzdana has received fees for participating in review activities from the Global Influenza Programme, WHO, and J. Enstone has received consultancy fees from GlaxoSmithKline.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

1
Meijer
A
Lackenby
A
Hungnes
O
, et al.  . 
Oseltamivir-resistant influenza virus A (H1N1), Europe, 2007–08 season
Emerg Infect Dis
 , 
2009
, vol. 
15
 (pg. 
552
-
60
)
2
Blyth
CC
Kelso
A
McPhie
KA
, et al.  . 
The impact of the pandemic influenza A(H1N1) 2009 virus on seasonal influenza A viruses in the southern hemisphere, 2009
Euro Surveill
 , 
2010
, vol. 
15
  
pii = 19631
3
Salisbury
D
Ramsay
M
Noakes
K
Influenza (updated January 2011)
Immunisation against infectious disease
 , 
2006
London
The Stationary Office
(pg. 
185
-
208
4
Beck
CR
McKenzie
BC
Hashim
AB
, et al.  . 
Influenza vaccination for immunocompromised patients: systematic review and meta-analysis from a public health policy perspective
PLoS One
 , 
2011
, vol. 
6
 pg. 
e29249
 
5
Beck
C
McKenzie
B
Hashim
A
Nguyen-Van-Tam
J
Clinical effectiveness of influenza vaccination for immunocompromised patients: a systematic review and meta-analysis
CRD Register
 , 
2011
 
CRD42011001226
6
Downs
SH
Black
N
The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions
J Epidemiol Community Health
 , 
1998
, vol. 
52
 (pg. 
377
-
84
)
7
Higgins
J
Green
S
Cochrane handbook for systematic reviews of interventions
 , 
2009
 
Version 5.0.2 (updated September 2009).The Cochrane Collaboration, Chichester, UK
8
West
S
King
V
Carey
T
, et al.  . 
Systems to rate the strength of scientific evidence
2002
 
Evidence report/technology appraisal number 47. Agency for Healthcare Research and Quality, Rockville, MD,
9
Committee for Proprietary Medicinal Products
Note for guidance on harmonisation of requirements for influenza vaccines
1997
London, United Kingdom
European Agency for the Evaluation of Medicinal Products
 
CPMP/BWP/214/96
10
ICH Expert Working Group
ICH harmonised tripartite guideline: guideline for good clinical practice E6(R1)
 
11
Centre for Reviews and Dissemination
Systematic reviews: CRD's guidance for undertaking reviews in health care
 , 
2008
York, United Kingdom
Centre for Reviews and Dissemination, University of York
12
Ranieri
R
Veronelli
A
Santambrogio
C
Pontiroli
AE
Impact of influenza vaccine on response to vaccination with pneumococcal vaccine in HIV patients
AIDS Res Human Retroviruses
 , 
2005
, vol. 
21
 (pg. 
407
-
9
)
13
Tasker
SA
Treanor
JJ
Paxton
WB
Wallace
MR
Efficacy of influenza vaccination in HIV-infected persons: a randomized, double-blind, placebo-controlled trial
Ann Intern Med
 , 
1999
, vol. 
131
 (pg. 
430
-
3
)
14
Anema
A
Mills
E
Montaner
J
Brownstein
JS
Cooper
C
Efficacy of influenza vaccination in HIV-positive patients: a systematic review and meta-analysis
HIV Med
 , 
2008
, vol. 
9
 (pg. 
57
-
61
)
15
Atashili
J
Kalilani
L
Adimora
AA
Efficacy and clinical effectiveness of influenza vaccines in HIV-infected individuals: a meta-analysis
BMC Infect Dis
 , 
2006
, vol. 
6
 pg. 
138
 
16
Yamanaka
H
Teruya
K
Tanaka
M
, et al.  . 
Efficacy and immunologic responses to influenza vaccine in HIV-1-infected patients
J Acquir Immune Defic Syndr
 , 
2005
, vol. 
39
 (pg. 
167
-
73
)
17
Gunthard
HF
Wong
JK
Spina
CA
, et al.  . 
Effect of influenza vaccination on viral replication and immune response in persons infected with human immunodeficiency virus receiving potent antiretroviral therapy
J Infect Dis
 , 
2000
, vol. 
181
 (pg. 
522
-
31
)
18
Rosok
B
Voltersvik
P
Bjerknes
R
Axelsson
M
Haaheim
LR
Asjo
B
Dynamics of HIV-1 replication following influenza vaccination of HIV+ individuals
Clin Exp Immunol
 , 
1996
, vol. 
104
 (pg. 
203
-
7
)
19
Banic
S
Koren
S
Tomazic
J
, et al.  . 
Influenza vaccination of human immunodeficiency virus 1-infected patients receiving antiretroviral therapy
Acta Virol
 , 
2001
, vol. 
45
 (pg. 
39
-
44
)
20
Vigano
A
Bricalli
D
Trabattoni
D
, et al.  . 
Immunization with both T cell-dependent and T cell-independent vaccines augments HIV viral load secondarily to stimulation of tumor necrosis factor alpha
AIDS Res Human Retroviruses
 , 
1998
, vol. 
14
 (pg. 
727
-
34
)
21
Staprans
SI
Hamilton
BL
Follansbee
SE
, et al.  . 
Activation of virus replication after vaccination of HIV-1-infected individuals
J Exp Med
 , 
1995
, vol. 
182
 (pg. 
1727
-
37
)
22
O'Brien
WA
Grovit-Ferbas
K
Namazi
A
, et al.  . 
Human immunodeficiency virus-type 1 replication can be increased in peripheral blood of seropositive patients after influenza vaccination
Blood
 , 
1995
, vol. 
86
 (pg. 
1082
-
9
)
23
Tasker
SA
O'Brien
WA
Treanor
JJ
, et al.  . 
Effects of influenza vaccination in HIV-infected adults: a double-blind, placebo-controlled trial
Vaccine
 , 
1998
, vol. 
16
 (pg. 
1039
-
42
)
24
Borella
L
Webster
RG
The immunosuppressive effects of long-term combination chemotherapy in children with acute leukemia in remission
Cancer Res
 , 
1971
, vol. 
31
 (pg. 
420
-
6
)
25
Esposito
S
Cecinati
V
Scicchitano
B
, et al.  . 
Impact of influenza-like illness and effectiveness of influenza vaccination in oncohematological children who have completed cancer therapy
Vaccine
 , 
2010
, vol. 
28
 (pg. 
1558
-
65
)
26
de Lavallade
H
Garland
P
Sekine
T
, et al.  . 
Repeated vaccination is required to optimize seroprotection against H1N1 in the immunocompromised host
Haematologica
 , 
2011
, vol. 
96
 (pg. 
307
-
14
)
27
Mazza
JJ
Yale
SH
Arrowood
JR
, et al.  . 
Efficacy of the influenza vaccine in patients with malignant lymphoma
Clin Med Res
 , 
2005
, vol. 
3
 (pg. 
214
-
20
)
28
Schafer
AI
Churchill
WH
Ames
P
Weinstein
L
The influence of chemotherapy on response of patients with hematologic malignancies to influenza vaccine
Cancer
 , 
1979
, vol. 
43
 (pg. 
25
-
30
)
29
Rapezzi
D
Sticchi
L
Racchi
O
Mangerini
R
Ferraris
AM
Gaetani
GF
Influenza vaccine in chronic lymphoproliferative disorders and multiple myeloma
Eur J Haematol
 , 
2003
, vol. 
70
 (pg. 
225
-
30
)
30
Shahgholi
E
Ehsani
MA
Salamati
P
Maysamie
A
Sotoudeh
K
Mokhtariazad
T
Immunogenicity of trivalent influenza vaccine in children with acute lymphoblastic leukemia during maintenance therapy
Pediatr Blood Cancer
 , 
2010
, vol. 
54
 (pg. 
716
-
20
)
31
Antonio
R
Delogu
G
Sali
M
, et al.  . 
Efficacy of influenza A(H1N1v) vaccination in children with cancer
Pediatr Blood Cancer
 , 
2010
, vol. 
55
 pg. 
973
 
32
Xu
Y
Methuku
N
Coimbatore
P
, et al.  . 
A prospective study of the immunogenicity of inactivated influenza A (H1N1) 2009 monovalent vaccine in patients who have solid or hematologic malignancies
 
48th Annual Meeting of the Infectious Diseases Society of America; 21–24 October 2010; Vancouver, Canada Abstract 5075, poster LB-13
33
Yalçin
SS
Kondolot
M
Albayrak
N
, et al.  . 
Serological response to influenza vaccine after hematopoetic stem cell transplantation
Ann Hematol
 , 
2010
, vol. 
89
 (pg. 
913
-
8
)
34
Thibault
F
Noble
CP
Morelon
E
, et al.  . 
Safety and efficacy of an inactivated, unadjuvanted vaccine adainst the novel influenza A variant (H1N1v) in renal transplant recipients (Transfluvac study)
 
In: Options for the Control of Influenza VII; 3–7 September 2010; Hong Kong, SAR, China Poster P573
35
Madan
RP
Tan
M
Fernandez-Sesma
A
, et al.  . 
A prospective, comparative study of the immune response to inactivated influenza vaccine in pediatric liver transplant recipients and their healthy siblings
Clin Infect Dis
 , 
2008
, vol. 
46
 (pg. 
712
-
8
)
36
Dopp
JM
Wiegert
NA
Moran
JJ
, et al.  . 
Effect of annual influenza immunization on antibody response in lung transplant patients
Prog Transplant
 , 
2009
, vol. 
19
 (pg. 
153
-
9
)
37
Scharpe
J
Evenepoel
P
Maes
B
, et al.  . 
Influenza vaccination is efficacious and safe in renal transplant recipients
Am J Transplant
 , 
2008
, vol. 
8
 (pg. 
332
-
7
)
38
Del Porto
F
Lagana
B
Biselli
R
, et al.  . 
Influenza vaccine administration in patients with systemic lupus erythematosus and rheumatoid arthritis. Safety and immunogenicity
Vaccine
 , 
2006
, vol. 
24
 (pg. 
3217
-
23
)
39
Holvast
A
van Assen
S
de Haan
A
, et al.  . 
Effect of a second, booster, influenza vaccination on antibody responses in quiescent systemic lupus erythematosus: an open, prospective, controlled study
Rheumatology
 , 
2009
, vol. 
48
 (pg. 
1294
-
9
)
40
Lu
CC
Wang
YC
Lai
JH
Lee
TSH
Lin
HT
Chang
DM
A/H1N1 influenza vaccination in patients with systemic lupus erythematosus: safety and immunity
Vaccine
 , 
2011
, vol. 
29
 (pg. 
444
-
50
)
41
Wiesik-Szewczyk
E
Romanowska
M
Mielnik
P
, et al.  . 
Anti-influenza vaccination in systemic lupus erythematosus patients: an analysis of specific humoral response and vaccination safety
Clin Rheumatol
 , 
2010
, vol. 
29
 (pg. 
605
-
13
)
42
Zykov
MP
Rafaelskaya
TI
Volkon
VM
Knoring
BE
Shenderova
RI
Baymakova
EA
Influenza virus vaccination of patients with pulmonary tuberculosis
Problemy Tuberkuleza
 , 
1980
, vol. 
58
 (pg. 
3
-
5
)
43
Kostyanaya
IE
Meisner
AF
Aksenova
VA
Baturo
AP
Experience in administration of the vaccines “Pneumo 23” and “VAXIGRIP” among children at risk infected with Mycobacterium tuberculosis
Bulleten Vakzinazija
 , 
2002
, vol. 
1
  
[epub]. http://medi.ru/doc/15b1906.htm Accessed 8 August 2012.
44
Hanania
NA
Sockrider
M
Castro
M
, et al.  . 
Immune response to influenza vaccination in children and adults with asthma: effect of corticosteroid therapy
J Allergy Clin Immunol
 , 
2004
, vol. 
113
 (pg. 
717
-
24
)
45
Kubiet
MA
Gonzalez-Rothi
RJ
Cottey
R
Bender
BS
Serum antibody response to influenza vaccine in pulmonary patients receiving corticosteroids
Chest
 , 
1996
, vol. 
110
 (pg. 
367
-
70
)
46
Nicholson
K
Nguyen-Van-Tam
J
Ahmed
A
, et al.  . 
Randomised placebo-controlled crossover trial on effect of inactivated influenza vaccine on pulmonary function in asthma
Lancet
 , 
1998
, vol. 
351
 (pg. 
326
-
31
)
47
Park
CL
Frank
AL
Sullivan
M
Jindal
P
Baxter
BD
Influenza vaccination of children during acute asthma exacerbation and concurrent prednisone therapy
Pediatrics
 , 
1996
, vol. 
98
 (pg. 
196
-
200
)
48
Reid
DW
Bromly
CL
Stenton
SC
Hendrick
DJ
Bourke
SJ
A double-blind placebo-controlled study of the effect of influenza vaccination on airway responsiveness in asthma
Respir Med
 , 
1998
, vol. 
92
 (pg. 
1010
-
1
)
49
Fiore
AE
Uyeki
TM
Broder
K
, et al.  . 
Prevention and control of influenza with vaccines: recommendations of the Advisory Committee on Immunization Practices (ACIP), 2010
MMWR Recomm Rep
 , 
2010
, vol. 
59
 (pg. 
1
-
62
)

Author notes

Presented in part: Expert consultation convened by World Health OrganizationDevelopment of a Standard Guideline on Clinical Management of Influenza Virus Infection, 8–9 June 2011, Geneva, Switzerland; Fourth European Scientific Working Group on Influenza Conference, September 2011.
a
Members of the UNIIC Study Group are listed in the Notes.