Decreased Antibody Response to Influenza Vaccination in Kidney Transplant Recipients: A Prospective Cohort Study

Results 

Baseline Characteristics 

A total of 107 healthy control individuals and 59 kidney transplant recipients were enrolled in the study during October and November 2006; of these, 106 control and 53 transplant recipient participants completed the second follow-up visit (99% and 90%, respectively; Fig 1). No difference in the ages of the 2 groups was seen (control mean age, 41.0 ± 12 years; transplant recipient mean age, 44.0 ± 10.3 years). Other demographics varied, with more women in the healthy control group and more African Americans in the transplant recipient group. As expected, measures of kidney function were significantly different for the 2 groups (P < 0.001). Both groups had similar high rates of reported influenza vaccination in prior years. Baseline seroprotection, represented as a titer of 1:32 or greater before receiving the influenza vaccination, was noted in 45% or greater of both groups for all 3 influenza types (Table 1).

View Large Image Download to PowerPoint

Figure 1. Study enrollment and follow up.

Table 1. Characteristics of the 159 Participants

		Controls (n = 106)	Transplant Recipients (n = 53)	P
	Age (y)	41.0±12.1	44.0±10.3	0.1
	Women	72(67.9)	20(37.7)	<0.001
	Race⁎			0.003
	White	92(86.8)	42(79.2)
	Black	4(3.8)	10(18.9)
	Asian	8(7.5)	0(0)
	American Indian or Alaskan native	2(1.9)	1(1.9)
	Serum creatinine(mg/dL)	0.9±0.2	1.5±0.6	<0.001
	Glomerular filtration rate (mL/min/1.73 m2)	85.1±17.4	59.8±19.7	<0.001
	Previous influenza vaccination†	89(84.0)	50(94.3)	0.2
	Baseline seroprotection‡
	A/H1N1	58(54.7)	32(60.4)	0.5
	A/H3N2	70(66.0)	24(45.3)	0.01
	B	62(58.5)	33(62.3)	0.7

Note: Values expressed as mean ± SD or number (percent). Conversion factors for units: serum creatinine in mg/dL to μmol/L, ×88.4; glomerular filtration rate in mL/min/1.73 m2 to mL/s/1.73 m2, ×0.01667.

⁎ Race was self-reported. † Previous influenza vaccination was self-reported. ‡ Seroprotection defined as antibody titer of 1:32 or greater.

As listed in Table 2, diabetes mellitus was the most common cause of end-stage renal disease (26.4%) in transplant recipient participants, and about half had a cadaveric donor source. Median time since transplantation was 307 days (interquartile range, 103.5 to 837). Following the study design, all participants were administered tacrolimus-based regimens, with two-thirds of participants administered either tacrolimus/mycophenolate mofetil or tacrolimus/mycophenolate mofetil/prednisone protocols. About 30% of participants were on antiviral prophylaxis therapy, reflecting a significant portion of participants in the early posttransplantation period.

Table 2. Characteristics of the Transplant Recipient Group

		Total (N = 53)	Transplantation < 6 mo (n = 19)	Transplantation > 6 mo (n= 34)	P
	Age (y)	44.0±10.3	46.4±9.3	42.7±10.7	0.3
	Women	20(37.7)	8(42.1)	12(35.3)	0.6
	Race⁎				0.7
	White	42(79.2)	15(78.9)	27(79.4)
	Black	10(18.9)	4(21.1)	6(17.6)
	Asian	0(0)	0(0)	0(0)
	American Indian or Alaskan native	1(1.9)	0(0)	1(2.9)
	Serum creatinine (mg/dL)	1.5±0.6	1.4±0.6	1.5±0.7	0.4
	Glomerular filtration rate (mL/min/1.73 m2)	59.8±19.7	61.8±18.8	58.6±20.3	0.9
	Previous influenza vaccination†	50(94.3)	18(94.7)	32(94.1)	0.9
	Baseline seroprotection‡
	A/H1N1	32(60.4)	11(57.9)	21(61.8)	0.8
	A/H3N2	24(45.3)	8(42.1)	16(47.1)	0.7
	B	33(62.3)	12(63.2)	21(61.8)	0.9
	Type of transplant				0.2
	Cadaveric	27(50.9)	12(63.2)	15(44.1)
	Living	26(49.1)	7(36.8)	19(55.9)
	Time since transplantation (d)	307(104-837)	70(36-177)	650(312-1,137)	<0.001
	Cause of kidney failure				0.2
	Diabetes	14(26.4)	7(36.8)	7(20.6)
	Glomerulonephritis	12(22.6)	3(15.8)	9(26.5)
	Polycystic kidney disease	10(18.9)	6(31.6)	4(11.8)
	Hypertension	9(17.0)	2(10.5)	7(20.6)
	Obstruction	5(9.4)	0(0)	5(14.7)
	Other/unknown	3(5.7)	1(5.3)	2(5.8)
	First transplant	47(88.7)	17(89.5)	30(88.2)	0.9
	Prior rejection episode	8(15.1)	0(0)	8(23.5)	0.02
	Induction agent				0.3
	Antithymocyte antiglobulin	33(62.3)	14(73.7)	19(55.9)
	Basiliximab	12(22.6)	4(21.0)	8(23.5)
	Alemtuzumab	3(5.7)	1(5.3)	2(5.9)
	Other	5(9.4)	0(0)	5(14.7)
	Immunosuppression regimen				<0.001
	Tacrolimus/MMF	18(34.0)	15(78.9)	3(8.8)
	Tacrolimus/MMF/prednisone	17(32.1)	2(10.5)	15(44.1)
	Tacrolimus/MPA	4(7.5)	1(5.3)	3(8.8)
	Tacrolimus/MPA/prednisone	6(11.3)	1(5.3)	5(14.7)
	Tacrolimus/prednisone	7(13.2)	0(0)	7(20.6)
	Other	1(1.9)	0(0)	1(2.9)
	Tacrolimus total daily dose (mg)	5.7±3.1	5.5±2.2	5.9±3.6	0.9
	Tacrolimus trough level (ng/mL)	6.9±2.1	7.4±1.9	6.6±2.1	0.1
	Prednisone use	30(56.6)	3(15.8)	27(79.4)	<0.001
	MMF/MPA use	44(83)	19(100)	25(73.5)	0.01
	MMF total daily dose(mg)§	1,600±497	1,765±437	1,444±511	0.6
	MPA total daily dose (mg)∥	1,008±372	1,080±509	990±373	0.8
	Antiviral prophylaxis	16(30.2)	13(68.4)	3(8.8)	<0.001

Note: Values expressed as mean±SD, median (25th to 75th percentiles), or number (percent). P refers to comparison between those who underwent transplantation less than 6 months and more than 6 months before. Conversion factors for units: serum creatinine in mg/dL to μmol/L, ×88.4; glomerular filtration rate in mL/min/1.73 m2 to mL/s/1.73 m2, ×0.01667. Abbreviations: MMF, mycophenolate mofetil; MPA, mycophenolic acid.

⁎ Race was self-reported. † Previous influenza vaccination was self-reported. ‡ Seroprotection defined as antibody titer of 1:32 or greater. § Thirty-five participants were administered MMF, 17 underwent transplantation less than 6 months before, and 18 underwent transplantation greater than 6 months before. ∥ Ten participants were administered MPA, 2 underwent transplantation less than 6 months before, and 8 underwent transplantation greater than 6 months before.

Antibody Titers 

Geometric mean titers are shown in Fig 2. No significant differences in prevaccination geometric mean titers for the A/H1N1 and B types were observed between the control and transplant recipient groups, with the difference for A/H3N2 of borderline significance (36.7 for controls, 23.8 for transplant recipients; P = 0.05). Postvaccination titers increased significantly from baseline for all 3 influenza types within both groups. Mean change in prevaccination to postvaccination titers was significantly different between the control and transplant recipient groups for all 3 types.

View Large Image Download to PowerPoint

Figure 2. Serological response to the 2006-2007 inactivated influenza vaccine expressed as geometric mean titers. For the control group: A/H1N1 prevaccination (pre), 33.6, postvaccination (post), 108.9; A/H3N2 pre, 36.7, post, 180.4; B pre, 36.1, post, 130.1. For the transplant recipient group, A/H1N1 pre, 34.2, post, 53.2; A/H3N2 pre, 23.8, post, 55.7; B pre, 40, post, 80.2.

Primary Outcome: Seroresponse and Seroprotection 

A smaller proportion of the transplant recipient group compared with the healthy control group developed the primary outcomes of seroresponse or seroprotection for all 3 influenza types at 1 month postvaccination (Table 3). Differences were statistically significant for only influenza type A/H3N2, for which the transplant recipient group had 0.31 decreased odds of developing seroresponse (95% confidence interval, 0.16 to 0.62; P = 0.001) and 0.22 decreased odds of developing seroprotection (95% confidence interval, 0.09 to 0.53; P = 0.001) compared with healthy controls (Figure 3, Figure 4). These results were unchanged after adjusting for age, sex, and race (Table 4). On univariate analysis, neither basic demographic factors nor serum creatinine level were strong predictors across all outcomes (Table S1; provided as online supplementary material available with this article at www.ajkd.org).

Table 3. Percentages and Odds Ratios of Developing Antibody Response to the 2006-2007 Influenza Vaccine in All Participants

Note: Model adjusted for control versus transplant participant, age, and baseline serum creatinine level. *n = 19 for transplant participants less than 6 months from time of transplantation, n = 34 for transplant participants greater than 6 months from time of transplantation.

View Large Image Download to PowerPoint

Figure 3. Odds ratios (95% confidence intervals) for the primary outcome of seroresponse (4-fold increase in antibody titer).

View Large Image Download to PowerPoint

Figure 4. Odds ratios (95% confidence intervals) for the primary outcome of seroprotection (antibody titer ≥ 1:32).

Table 4. Unadjusted and Adjusted Odds Ratios of Kidney Transplant Recipients Developing the Outcomes of Seroresponse or Seroprotection Compared With Controls

		Seroresponse Odds Ratio (95% confidence interval)			Seroprotection Odds Ratio (95% confidence interval)
		A/H1N1	A/H3N2	B	A/H1N1	A/H3N2	B
	Group	0.62(0.29-1.33)	0.31(0.16-0.62)	0.60(0.31-1.12)	0.47(0.22-1.03)	0.22(0.09-0.53)	0.54(0.18-1.57)
	Group + age	0.69(0.32-1.50)	0.32(0.16-0.64)	0.65(0.33-1.23)	0.51(0.23-1.13)	0.22(0.09-0.54)	0.58(0.20-1.71)
	Group + sex⁎	0.51(0.23-1.13)	0.29(0.14-0.61)	0.63(0.31-1.27)	0.48(0.21-1.08)	0.29(0.11-0.74)	0.39(0.13-1.23)
	Group + race	0.62(0.29-1.36)	0.31(0.16-0.62)	0.60(0.30-1.18)	0.47(0.22-1.03)	0.21(0.09-0.52)	0.52(0.18-1.53)
	Group + serum creatinine	0.89(0.35-2.25)	0.28(0.12-0.66)	0.47(0.20-1.09)	0.48(0.19-1.22)	0.28(0.10-0.80)	0.43(0.12-1.55)
	Group + previous vaccination	0.66(0.31-1.44)	0.31(0.16-0.62)	0.62(0.31-1.23)	0.48(0.22-1.04)	0.18(0.07-0.46)	0.52(0.18-1.54)

Note: Group is kidney transplant recipients versus healthy controls.

⁎ Reference category for sex is women.

One-third of the transplant recipient participants were 6 months or less from the date of kidney transplantation, a period of more intense immunosuppression. Therefore, stratified analyses were conducted for kidney transplant recipient participants less than or greater than 6 months (defined as < or > 183 days) from the time of transplantation compared with the healthy control group. Of 53 transplant recipient participants, 19 were less than 6 months and 34 were greater than 6 months from transplantation. The 2 transplant recipient groups did not differ by age, sex, race, baseline kidney function, cadaveric versus living donor source, induction agent, prior influenza vaccination, or baseline seroprotection. More prednisone use was seen in the group greater than 6 months posttransplantation (Table 2). For the outcome seroresponse, kidney transplant recipient participants within 6 months of transplantation were significantly less likely than healthy controls to develop a 4-fold increase in titers for all 3 influenza types. Transplant recipients more than 6 months from transplantation also were significantly less likely than controls to develop seroresponse for type A/H3N2 (Table 3; Fig 3). Seroprotection results were similar for the A subtypes, with those less than 6 months from transplantation having decreased odds of developing seroprotection compared with healthy controls. No difference was seen for B (Fig 4).

Subgroup analysis was performed considering only participants without baseline seroprotection because participants with baseline seroprotection would already have met the outcome of seroprotection and also might influence the results of seroresponse. By selecting these participants, the new cohort size was 69 for A/H1N1, 65 for A/H3N2, and 64 for B. The odds of susceptibles achieving a protective titer ranged from 54% to 85% less in transplant recipient participants, supporting results of the full cohort (Table 5).

Table 5. Percentages and Odds Ratios of Developing Antibody Response to the 2006-2007 Influenza Vaccine in Participants Without Baseline Seroprotection

Note: Model adjusted for control versus transplant participant and age. *n = 48 for controls and n = 21 (n = 8 < 6 months, n = 13 > 6 months) for transplant recipients for A/H1N1. †n = 36 for controls and n = 29 (n = 11 < 6 months, n = 18 > 6 months) for transplant recipients for A/H3N2. ‡n = 44 for controls and n = 20 (n = 7 < 6 months, n = 13 > 6 months) for transplant recipients for B.

As noted, baseline kidney function differed between the 2 groups. To understand how kidney function might influence vaccine response, multivariate logistic regression was performed. Covariates included were group (transplant recipient versus control), age, and serum creatinine level. Overall results were unchanged (Table 3). Because of the almost nonoverlapping distribution of group and serum creatinine level, strong colinearity between these 2 covariates produced too much instability for the model to be used for the smaller cohort of participants with no baseline seroprotection. This cohort was adjusted for group and age (Table 5).

Serum creatinine levels in transplant recipient participants at the 1-month follow-up visit were not significantly different from baseline values. No adverse events were reported by any study participant, consistent with the known safety profile of the inactivated influenza vaccine.

Discussion 

In this prospective cohort study, we show that the proportion of participants developing a 4-fold increase in titer or a postvaccination titer greater than 1:32 to the 2006-2007 inactivated influenza vaccine was decreased in kidney transplant recipients compared with healthy controls. Although the overall difference was statistically significant for only the A/H3N2 type, participants less than 6 months posttransplantation had significantly decreased antibody responses for all 3 influenza types compared with healthy controls. In addition, controlling for the presence of baseline seroprotection magnified the difference observed between the transplant recipient and healthy control groups. These results may be clinically important to the practice of influenza vaccination in kidney transplant recipients on current immunosuppression protocols including tacrolimus. A more definitive study with a larger sample size would be useful for generalizing these results.

Tacrolimus is a type of calcineurin inhibitor that blocks T-cell lymphokine production and inhibits B-cell activation and proliferation, putting participants at greater risk of bacterial and viral infection20 and potentially a decreased response to vaccines dependent on these functions. Previous studies evaluating cyclosporine, an older type of calcineurin inhibitor no longer used as commonly for post–kidney transplantation immunosuppression, suggested it caused decreased antibody response to influenza vaccine. Two studies compared kidney transplant recipients treated with cyclosporine versus azathioprine, and both found the cyclosporine groups had a decreased response to influenza vaccine compared with controls.9, 10 We selected participants using only tacrolimus-based regimens because tacrolimus is part of the most widely used initial regimens for kidney transplantation in the United States. Tacrolimus formed 82% of all regimens for 2006 compared with 12% for cyclosporine.21 By focusing on this drug, we also minimized our chance of seeing no significant difference based on confounders that may have been introduced by considering multiple different regimens at once.

The immediate period posttransplantation is one of intense immunosuppression because of the use of induction immunosuppression agents at time of surgical transplantation and greater doses of immunosuppressive medication in the first several months. The immunosuppressive effects supplied by an induction agent depend on the agent used, but can last months.22, 23 The schedule for tapering the dose of immunosuppression varies by center, but in general, as practiced at our institution, doses usually are tapered after 6 months, providing an arbitrary, but clinically reasonable, cutoff point for greater or lesser immunosuppression for a general kidney transplant recipient. Accordingly, when stratifying the transplant recipient group by less or greater than 6 months from the time of transplantation, we were able to detect significantly poorer antibody responses to the influenza vaccine in transplant recipients less than 6 months posttransplantation compared with healthy controls. These findings retained statistical significance, although the transplant recipient groups contained smaller numbers of participants once divided, meaning we had less power to find such differences.

At our institution, influenza vaccination routinely is offered to recipients who are 1 month or more posttransplantation, which may vary from other centers that choose to wait longer. Our cohort, with a median time from transplantation of 307 days, has a considerably decreased time from transplantation compared with the 2 most recent published studies about kidney transplant recipients and influenza vaccination. In these studies, participants less than 6 months from transplantation were excluded, and time from transplantation was 53 months (mean) in 1 study and 6.3 years (median) in the other.24, 25 Our finding suggests the vaccine may not work as well in these early posttransplantation participants. As a result, clinicians should have a greater suspicion for susceptibility to influenza despite vaccination in this group. One also could consider not administering the vaccine until a recipient is at least 6 months posttransplantation. However, although those who underwent transplantation less than 6 months before do not respond as well as healthy controls, some individuals have a response and are protected, raising the question of the cost-benefit of giving the vaccine earlier. This group could be studied further to see whether a second vaccine dose, such as administered to children, would help boost antibody response.26, 27

The difference in antibody response to A/H3N2 versus the other influenza types is not unusual. Responses to different components of the vaccine may vary based on prior exposures to circulating wild-type viruses and previous vaccination.28, 29, 30 These exposures also resulted in several cohort members having the presence of protective antibodies greater than 1:32. Clinically, the proportions of participants without baseline seroprotection are interesting to examine because they are the subgroup at risk of influenza infection during the concurrent season. As a result, we analyzed these participants separately and found results similar to the larger cohort. As determined in other studies, the presence of baseline seroprotection before vaccination results in less robust vaccine responses.8, 25, 28

Our findings differ from the main findings of 2 recent reports that examined cohorts of kidney transplant recipients and found that seroresponse and seroprotection postvaccination were similar between transplant recipients and healthy controls.24, 25 In addition to differences in time from transplantation, as discussed, these studies differ from the present study in types of immunosuppression used. In the study by Scharpé et al,25 the transplant recipient group was administered a variety of immunosuppressants, whereas our study considers only transplant recipients administered tacrolimus-based regimens because we were concerned with defining responses in kidney transplant recipient participants on the most common type of calcineurin inhibitor in use. Keshtkar-Jahromi et al24 compared vaccine responses in participants administered mycophenolate mofetil versus azathioprine in addition to prednisolone and cyclosporine.

Several limitations of our study exist. First, because everyone received the vaccine, it was not possible to randomly assign participants. One way to address this in future studies is to match the participants for certain variables, such age, sex, or race, to help minimize confounding. Second, although the proportion of responders in the overall transplantation group was always less for each influenza type compared with healthy controls, our primary outcome measures of seroresponse and seroprotection were not always statistically significant. Given this trend, a larger cohort might provide increased power to find this difference statistically significant. Third, we were unable to perform extensive multivariate analysis because of small numbers of events/nonevents. However, we were able to control for some important confounders through limitations of the inclusion criteria, stratified analysis, and subgroup analysis. One potential confounder is decreased kidney function. Previous studies have shown diminished response to influenza vaccination in participants with chronic kidney disease, particularly in dialysis patients.31, 32 In kidney transplant recipients, older studies have conflicted on the role of immunosuppression versus decreased kidney function as a cause for impaired vaccine response.5, 33 Because the transplant recipient and control groups had an almost nonoverlapping distribution of serum creatinine values, adjustment for serum creatinine level in the multivariable logistic regression model introduced instability to the model because of colinearity between the presence of the transplant and serum creatinine level. However, overall results for the full cohort were unchanged when adjusted for serum creatinine level. The addition of a concurrent group of patients with chronic kidney disease with kidney function similar to the transplant recipient group should be explored in future studies.

In summary, we found that the proportion of transplant recipient participants with antibody response to the inactivated influenza vaccine was consistently less compared with controls for all 3 influenza types, but this reaches statistical significance for only A/H3N2. When participants less than 6 months from the time of transplantation were considered, this group had significantly decreased responses to the vaccine compared with healthy controls, especially for the A subtypes. Future studies comparing transplant recipients with patients with chronic kidney disease with similar kidney function or using influenza vaccines with booster doses in participants within 6 months of transplantation may provide more comprehensive information in this area.