Risk of underlying chronic medical conditions for invasive pneumococcal disease in adults

Vaccine xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Vaccine journal homepage: www.elsevier.com/locate/vaccine

Risk of underlying chronic medical conditions for invasive pneumococcal disease in adults Roger Baxter a,⇑, Arnold Yee a, Laurie Aukes a, Vincenza Snow b, Bruce Fireman a, Bruce Atkinson b, Nicola P. Klein a a b

Northern California Kaiser Permanente Vaccine Study Center, Oakland, CA, USA Pfizer Vaccines, 500 Arcola Road, Collegeville, PA, USA

a r t i c l e

i n f o

Article history: Received 27 January 2016 Received in revised form 14 June 2016 Accepted 4 July 2016 Available online xxxx Keywords: Pneumococcal Streptococcus pneumoniae Invasive pneumococcal disease (IPD) Risk conditions

a b s t r a c t Purpose: In the United States, the 13-valent pneumococcal conjugate vaccine is recommended in persons P65 years of age, and persons 665 years of age with immunocompromising (IC) conditions. For invasive pneumococcal disease (IPD) prevention in those 665 with non-IC medical conditions, the 23-valent polysaccharide vaccine is recommended. This group is at higher risk of IPD than the general population, but the level of risk is not well-quantified. We estimated IPD risk by individual underlying medical conditions, and by total number of conditions, for persons P18 years of age. We calculated the relative risks (RR) of various medical conditions, comparing the incident IPD cases to the general study population, and used Poisson regression models to estimate an IPD RR, adjusting for other conditions. We also examined IPD incidence by number of conditions diagnosed in each calendar year, using a risk-stacking model. Results: Underlying medical conditions with the highest adjusted RR for IPD were chronic liver disease (RR 2.1, 95% CI 1.5–2.8) and chronic obstructive pulmonary disease (COPD; RR 2.1, 95% CI 1.8–2.5). IPD risk increased with increasing number of medical conditions: adjusted RR, 2.2 (95% CI 1.9–2.5) 1 condition, 2.9 (2.5–3.5) for 2 conditions, and 5.2 (4.4–6.1) for 3 conditions. Conclusions: For persons with a single, non-IC medical condition, IPD risk was twice that for the general KPNC population. Persons with multiple, non-IC chronic conditions exhibited increased IPD risk with each additional condition. Such information may inform discussions on recommendations for adult pneumococcal immunization and prevention. Ó 2016 Published by Elsevier Ltd.

1. Background The use of seven-valent pneumococcal conjugate vaccine (PCV7, PrevnarÒ) in infants and children since 2000 has greatly modified the landscape of invasive pneumococcal disease (IPD), across all age groups [1–3]. Data from numerous completed and ongoing studies have characterized the effects of the six additional serotypes contained in 13-valent pneumococcal conjugate vaccine (PCV13, Prevenar13TM) [4–6] since its introduction in 2010. In 2012, the Advisory Committee on Immunization Practices (ACIP) recommended use of PCV13 in immunocompromised adults aged 19 and above. In 2014, the ACIP expanded their recommendation to include routine use of PCV13 first in all adults aged P65 years [7] followed by a dose of the 23-valent pneumococcal polysaccharide vaccine (PPSV23, PneumovaxÒ). For those individuals over age ⇑ Corresponding author at: Kaiser Permanente Vaccine Study Center, 1 Kaiser Plaza, 16th Floor, Oakland, CA 94612, USA. E-mail address: [email protected] (R. Baxter).

65 who have already received PPSV23, the ACIP recommended that they also receive PCV13. PPSV23 alone is recommended as a single dose for immunocompetent persons under age 65 years with certain comorbidities which are not immunocompromising (IC), but which are known to carry an increased risk of IPD. The comorbidities for which the ACIP currently recommends PPSV23 are: chronic lung disease (including asthma and COPD), chronic heart disease, chronic liver disease (including cirrhosis), diabetes mellitus, alcoholism and smoking. These recommendations, while evidence-based, are complicated, and can be confusing for clinicians. Although the association between IC chronic medical conditions and IPD in adults is well-documented, the level of IPD risk associated with these chronic medical conditions is not well quantified. There are few population-based studies examining the relationship between certain comorbidities and the increased risk of IPD [3,8–11]. One study analyzed IPD risk in relation to the number of chronic conditions with which individual patients were diagnosed [10]. However, there remain fundamental, unanswered questions

http://dx.doi.org/10.1016/j.vaccine.2016.07.003 0264-410X/Ó 2016 Published by Elsevier Ltd.

Please cite this article in press as: Baxter R et al. Risk of underlying chronic medical conditions for invasive pneumococcal disease in adults. Vaccine (2016), http://dx.doi.org/10.1016/j.vaccine.2016.07.003

2

R. Baxter et al. / Vaccine xxx (2016) xxx–xxx

on the epidemiology of IPD in adults, and their IPD risk as a function of chronic, non-IC medical conditions. Studies designed to assess IPD in this adult at-risk population are essential for a more complete elucidation of the disease burden in this population, and can ultimately elevate awareness within the clinical community of where pneumococcal vaccines such as PCV13 will provide the greatest benefit for IPD prevention. The aim of this study therefore was to evaluate IPD risk driven by underlying medical conditions, among adults within a large integrated medical care organization. 2. Methods 2.1. Study population Kaiser Permanente Northern California (KPNC) is an integrated healthcare plan currently serving approximately 3.5 million members. KPNC members receive essentially all medical care at KPNC facilities. Inpatient, emergency department, and outpatient visit diagnoses, medications, laboratory testing, imaging procedures, and all other medical information is linked to a unique medical record number, which stays with a member for life. All microbiology specimens are processed within a single, central microbiology laboratory. Since 2006, Kaiser Permanente Vaccine Study Center personnel have collected all IPD specimens for adult KPNC members and sent them to Boston University for serotype testing. Adults P18 years of age were included in this study. The study years included 2008 through 2014. A minimum of 10 months of KPNC membership during the year prior to the onset date of IPD, or July 31 of the same year in the control population, was required for both the IPD cases and the general adult KPNC membership. 2.2. Data sources All data pertinent to this study were derived from internallycontrolled, captured, and stored KPNC databases. Age and race were derived from KPNC membership databases. IPD isolates were identified from the KPNC laboratory system. Diagnostic codes from the KPNC electronic medical record were used to identify chronic conditions including alcoholism, chronic liver disease, cigarette smoking, chronic kidney disease (CKD) as CKD3 or 4 (glomerular filtration rate [GFR] 15–59 mL/min), congestive heart failure, COPD, dementia, and stroke. Similarly, diagnostic codes were used to identify IC conditions including asplenia, cerebrospinal fluid (CSF) leak, congenital immunodeficiency, end-stage renal disease (ESRD; GFR < 15 or on dialysis), leukemia, lymphoma, generalized malignancy, myelofibrosis, multiple myeloma, nephrotic syndrome, sickle cell hemoglobinopathy, and cochlear implants. CSF leaks and cochlear implants were classified as immunocompromising because individuals with CSF leaks and cochlear implants are already recommended to receive PCV13. KPNC registries were used to identify asthma and diabetes mellitus, as well as the IC condition of HIV infection. KPNC pharmacy databases were used to identify members on IC therapy such as chemotherapy, and other IC medications, such as corticosteroids (CS), and rheumatologic TNF inhibitors. 2.3. Medical conditions criteria Underlying medical conditions were identified beginning 2 years prior to the date when the IPD occurred (index date). For persons without IPD, we used July 31 of the calendar year in which an age-matched IPD case occurred. Chemotherapies and IC medications were included if administered within the 90 days prior to the index date. Systemic corticosteroids were considered

immunocompromising if the total dose administered exceeded 700 mg in either of the two 90 day periods (i.e., approximately 6 months) just prior to the index date. 2.4. Study design We compared adult IPD cases with the entire adult KPNC membership for differences in age, sex, race, and chronic medical conditions. Poisson regression models were used to estimate an adjusted IPD relative risk (RR). We estimated the IPD RR separately for each chronic medical condition, compared with the entire KPNC health plan population. The models were stratified by year, age group, sex, and race, with each chronic and IC medical condition used as a predictor variable. In addition, the model included history of ever having received PPSV23. For the risk stacking model, we examined IPD risk in relation to the total number of conditions with which individuals were diagnosed in any one calendar year. Each KPNC member was assigned to a risk category depending on how many risk conditions they had during the year: 0 (healthy), 1, 2, or P3. Incidence rates were stratified by age group for each summarized risk category, including healthy, chronic conditions (one risk condition, two non-IC conditions, three or more non-IC conditions), and IC conditions. If a person had both chronic and IC conditions, they were classified as IC. Poisson regression models were stratified by year, age group, sex, and race, with each summarized risk category used as predictor variables. This stacking analysis also included history of ever having received PPV23. This study was approved by the KPNC Institutional Review Board. 3. Results We detected a total of 1549 IPD cases during the study period 2008 through 2014, across a total of 15,102,047 person-years in the adult KPNC membership. The IPD incidence rate per 100,000 person-years appeared to decline over the seven-year study period. The average adult IPD rate in the two years (2008–2009) prior to the introduction of pediatric PCV13 in 2010 was 11.8 cases per 100,000 adult members; this rate declined to an average of 9.2 cases per 100,000 adults over the last four study years (2011–2014). The IPD incidence rate was higher in each older age group (Fig. 1). Adults with IPD were more likely than the KPNC population to be P65 years of age, and less likely to be <50 years old. The highest IPD incidence rates were observed in Blacks, while the lowest rates were in Asians (Table 1). Known risk conditions for IPD, such as alcoholism, asthma, and liver disease, were much more common in IPD cases when compared to the KPNC population (Table 2). After adjusting for all IC and chronic conditions, and stratifying by calendar year, age group, sex, and race, all chronic medical conditions, except for stroke and dementia, were associated with a statistically-significant increase in IPD risk (Table 3). The highest observed RR for underlying non-IC conditions was 2.1, for both chronic liver disease and COPD. Cochlear implants and Sickle Cell hemoglobinopathy were excluded from the RR calculations due to an absence of cases. Stacking analysis: ‘‘Stacking” refers to estimating the additional IPD risk for persons having one or more chronic medical conditions, as noted in Methods. Since persons with IC conditions are already recommended to receive PCV13, we focused on stacking for chronic conditions. If a person had both chronic and IC conditions, they were classified as having both conditions for purposes of risk stacking analysis. The IPD RR for persons with chronic medical conditions increased from 2.2 (for persons with

Please cite this article in press as: Baxter R et al. Risk of underlying chronic medical conditions for invasive pneumococcal disease in adults. Vaccine (2016), http://dx.doi.org/10.1016/j.vaccine.2016.07.003

3

R. Baxter et al. / Vaccine xxx (2016) xxx–xxx

IPD Rate per 100,000 person-years

60.0 50.0 40.0 30.0 20.0 10.0 18 - < 50 years 50 - < 65 years 65 - < 80 years >= 80 years

2008 4.2 13.4 20.7 53.5

2009 6.1 13.1 22.0 40.5

2010 5.0 12.5 26.3 41.1

2011 3.8 11.2 20.3 47.0

2012 2.9 8.3 17.9 40.3

2013 2.5 8.3 19.7 42.4

2014 2.6 8.9 15.1 35.3

Fig. 1. Incidence of Invasive Pneumococcal Disease (IPD) per 100,000 Patient Years by Age Group, Kaiser Permanente Northern California (KPNC), 2008-2014.

Table 1 Age and race distribution of general KPNC adult membership, vs. Invasive Pneumococcal Disease (IPD) cases; and incidence rate of IPD by race, Kaiser Permanente, 2008–2014. KPNC membership (%)

IPD cases (%)

Age group in years 18 to <50 51 50 to <65 29 65 to <80 15 680 5

19 30 29 21

Race Asian Black Hispanic White Other

10 11 13 64 1

19 8 17 56 1

Incidence rate of IPD (per 100,000 person-years)

Proportion of membership vaccinated with PPSV23 (%)

Proportion of IPD cases vaccinated with PPSV23 (%)

5.5 14.8 8.1 11.8 13.9

20 24 18 28 23

55 57 64 60 64

IPD = invasive pneumococcal disease; KPNC = Kaiser Permanente Northern California. PPSV23 = 23-valent Pneumococcal conjugate vaccine.

Table 2 Percent of each chronic medical condition in IPD cases and in the KPNC adult membership and their unadjusted ratio, 2008–2014. Non-immunocompromising (Non-IC) conditions

Percent of IPD cases with condition (%)

Percent of adult KPNC membership with condition (%)

Unadjusted ratio

Alcoholism Asthma Chronic liver disease Cigarette smoking Chronic kidney disease stage 3 or 4 Congestive heart failure Chronic obstructive pulmonary disease Dementia Diabetes mellitus Stroke

5.0 19.7 3.0 30.9 14.3 11.6 14.1 1.9 25.8 1.3

1.8 7.5 0.4 10.3 3.5 1.4 1.6 0.3 10.5 0.3

2.8 2.6 7.5 3.0 4.1 8.3 8.8 6.3 2.5 4.3

Immunocompromising (IC) conditions

IPD cases

KPNC adult membership

Asplenia Cerebrospinal fluid leak Chemotherapy medications Cochlear implant Congenital immunodeficiency End stage renal disease Generalized malignancy HIV Immunocompromising medications Leukemia Lymphoma Myelofibrosis Multiple myeloma Nephrotic syndrome Sickle cell hemoglobinopathy

0.1 0.2 4.5 No cases 4.1 4.6 4.6 2.4 11.0 2.7 4.0 0.1 3.9 0.4 No cases

0.01 0.01 0.4 0.0 0.2 0.3 0.6 0.3 1.1 0.1 0.2 0.005 0.05 0.07 0.01

10 20 11.25 NEa 20.5 15.3 7.7 8 10 27 20 20 78 5.7 NEa

IPD = invasive pneumococcal disease; KPNC = Kaiser Permanente Northern California. a NE: not evaluable, as there were no IPD cases with these conditions.

Please cite this article in press as: Baxter R et al. Risk of underlying chronic medical conditions for invasive pneumococcal disease in adults. Vaccine (2016), http://dx.doi.org/10.1016/j.vaccine.2016.07.003

4

R. Baxter et al. / Vaccine xxx (2016) xxx–xxx

Table 3 Adjusted relative risk of IPD by age, race, pneumovax, sex, year and medical conditions, in comparison with the general population. These are all adjusted for in the Poisson Regression model. KPNC, 2008–2014.

Age group (ref = 18 to <50)

Race (ref = White)

Pneumovax (ref = NO) Sex (ref = Male) Year (ref = 2014)

50 to <65 years 65 to <80 years P80 years Asian Black Hispanic Other YES F 2008 2009 2010 2011 2012 2013

Adjusted relative risk

95% confidence interval

2.0

1.8–2.4

2.5

2.0–3.0

4.9 0.7 1.3 1.0 1.3 1.3 1.0 1.5 1.6 1.5 1.3 1.1 1.1

4.0–6.0 0.6–0.8 1.1–1.5 0.9–1.2 0.8–2.0 1.2–1.6 0.9–1.1 1.2–1.8 1.3–1.9 1.3–1.8 1.1–1.6 0.9–1.3 0.9–1.4

Non-immunocompromising (Non-IC) conditions Alcoholism 1.6 Asthma 1.8 Chronic liver disease 2.1 Cigarette smoking 1.6 Chronic kidney disease 1.2 stage 3,4 Congestive heart failure 1.5 Chronic obstructive 2.1 pulmonary disease Dementia 1.3 Diabetes mellitus 1.3 Stroke 1.0 Immunocompromising (IC) conditions Asplenia Cerebrospinal fluid leak Chemotherapy Congenital immunodeficiency End stage renal disease Generalized malignancy HIV Immunocompromising meds Leukemia Lymphoma Multiple myeloma Myelofibrosis Nephrotic syndrome IPD = invasive California.

pneumococcal

disease;

Race (ref = White)

Pneumovax (ref = NO) Sex (ref = Male) Year (ref = 2014)

1.2–1.7 1.8–2.5

3.7 1.8 7.0 2.5

2.8–4.7 1.4–2.3 4.9–9.7 2.0–3.0

4.3 3.9 11.9 1.5 0.6

3.1–5.9 2.9–5.1 8.9–15.7 0.1–6.8 0.2–1.2 Permanente

50 to <65 years 65 to <80 years P80 years Asian Black Hispanic Other YES F 2008 2009 2010 2011 2012 2013

Non-IC conditions (ref = Healthy)

Unadjusted Relative risk

1 Condition 2 Conditions P3 Conditions IC Conditions

3.0 5.5 13.4 24.0

Relative risk

95% confidence interval

1.9

1.7–2.2

2.4

2.0–2.9

4.6 0.7 1.3 1.0 1.3 1.3 0.9 1.5 1.6 1.6 1.3 1.1 1.1

3.8–5.5 0.6–0.8 1.1–1.5 0.8–1.1 0.8–2.0 1.1–1.5 0.8–1.0 1.3–1.8 1.3–1.9 1.3–1.9 1.1–1.6 0.9–1.3 0.9–1.4

Adjusted Relative risk 2.2 2.9 5.2 6.8

95% confidence interval 1.9–2.5 2.5–3.5 4.4–6.1 6.1–7.7

IC = immunocompromising; IPD = invasive pneumococcal disease.

0.9–1.9 1.2–1.5 0.6–1.6 0.7–12.9 1.6–16.7 1.6–2.7 1.8–3.1

KPNC = Kaiser

Age Group (ref = 18 to <50)

1.2–2.0 1.6–2.1 1.5–2.8 1.4–1.8 1.0–1.4

4.2 6.4 2.1 2.4

Table 4 Relative Risk of IPD, for one or multiple non-immunocompromising conditions, compared to the healthy population. The stacking model (Adjusted Relative Risk) controls for age, sex, year, PPSV23, and medical conditions, and adds an interaction term for patients with immunocompromising and non-immunocompromising conditions.

Northern

one condition) to 5.2 (for persons with P3 or conditions). Output from the same modeling yielded a RR of 6.8 in persons with IC conditions (Table 4).

4. Discussion While the elevated adult IPD risk driven by IC chronic medical conditions is well-described [3,12,13], the potential, relative impact on IPD associated with specific chronic conditions has not been fully explored. Relatively few studies have examined the relationship between the ten chronic conditions examined in this study and the potential for increased adult IPD risk. Chronic medical conditions, particularly diabetes, are prevalent in the U.S, and numerous persons have more than one chronic condition [14]. The aim of our observational study was to describe the IPD risk associated with individual and combined (i.e., stacked) chronic conditions among adults. For this purpose, we employed an

analytical model that controlled for age, sex, year of study, PPSV23 exposure, and medical conditions; it also incorporated an interaction term for patients with IC and chronic conditions. Using this model, the IPD RR for individuals with 1, 2, or P3 chronic conditions was 2.2, 2.9, and 5.2 times that for the general KPNC population, respectively. The KPNC system is advantageous for population-based studies of this type because it provides all patient health care, and complete access to patient medical records. As such, we were able to precisely define the longitudinal population-based IPD risk, as well as the potential impact of a number of chronic, underlying medical conditions, on IPD risk. Our study period included observation time both immediately prior to pediatric PCV13 implementation in 2010, as well as the first three years following pediatric PCV13 implementation in the US. The average IPD rate in the pre-PCV13 pediatric implementation years (2006–2009) was 11.8 cases per 100,000 adult members, declining to 9.2 cases per 100,000 after implementation (2011–2014). IPD incidence dynamics in our study varied among two distinct age groups within the KPNC membership. The first group, aged 18 to <65 years, comprised 80% of the study population, and accounted for 49% of the IPD cases; the age 50 to <65 subgroup (29% of the study population and 30% of IPD cases) had an IPD RR of 1.9–2.0, relative to the general KPNC population. The baseline

Table 5 Vaccination Coverage 2008–2014. Age group

IPD cohort (%)

KPNC membership (%)

18 to <50 years 50 to <65 years P65 years

24 39 86

4 19 82

IPD = invasive California.

pneumococcal

disease;

KPNC = Kaiser

Permanente

Northern

Please cite this article in press as: Baxter R et al. Risk of underlying chronic medical conditions for invasive pneumococcal disease in adults. Vaccine (2016), http://dx.doi.org/10.1016/j.vaccine.2016.07.003

R. Baxter et al. / Vaccine xxx (2016) xxx–xxx

5

RR IPD risk for this age subgroup, independent of medical conditions, is therefore double that of the general population. It is probable that persons within this age range currently have, or will develop, P1 medical conditions, including those which are non-IC, given the observed increase in such IPD-disposing conditions beginning after age 45 [15]. It is important to quantify IPD RR associated with these conditions, particularly for those who have non-IC conditions, because individuals with these conditions who are 665 years old are not currently recommended by the ACIP to receive PCV13. Persons 65 years and older comprised 20% of the study population, yet accounted for 50% of IPD; RRs for the age 65 to <80 and >80 year old subgroups were 2.5 and 4.9, respectively. PPSV23 vaccination rates for both the IPD cohort (86%) and KPNC general population (82%) aged P65 years (Table 5) were well above the estimated national average of 59.9% [16]. Half of all IPD cases were observed in PPSV23-vaccinated subjects, independent of concurrent medical conditions. Unadjusted risks in the stacking model were lowered considerably by adjusting for important variables, highlighting the benefit of a population-based model. In addition, our stacking analysis demonstrated that having multiple non-IC conditions markedly increases the subsequent risk of IPD.

group for which PPSV23 is the sole recommended pneumococcal vaccine. Concerns have arisen regarding both immediate and long-term immune responses to PPSV23, and have consequently led to questions on whether individuals with multiple—or even individual—non-IC conditions, should preferentially receive PCV13, in alignment with current recommendations for those with IC medical conditions who are P19 years of age, as well as those over 65. It is our hope that this analysis provides insights which will better inform these discussions.

4.1. Limitations

[1] Halasa NB, Grijalva CG, Arbogast PG, et al. Near complete elimination of the seven valent pneumococcal conjugate vaccine serotypes in tennessee. Pediatr Infect Dis J 2013;32:604–9. [2] Ampofo K, Pavia AT, Chris S, et al. The changing epidemiology of invasive pneumococcal disease at a tertiary children’s hospital through the 7-valent pneumococcal conjugate vaccine era: a case for continuous surveillance. Pediatr Infect Dis J 2012;31:228–34. [3] Muhammad RD, Oza-Frank R, Zell E, et al. Epidemiology of invasive pneumococcal disease among high-risk adults since the introduction of pneumococcal conjugate vaccine for children. Clin Infect Dis 2013;56:e59–67. [4] Lim GH, Wormsbecker AE, McGeer A, et al. Have changing pneumococcal vaccination programmes impacted disease in Ontario. Vaccine 2013;31 (24):2680–5. [5] Kaplan SL, Barson WJ, Lin PL, et al. Early trends for invasive pneumococcal infections in children after the introduction of the 13-valent pneumococcal conjugate vaccine. Pediatr Infect Dis J 2013;32:203–7. [6] Demczuk WH, Martin I, Griffith A, et al. Serotype distribution of invasive Streptococcus pneumoniae in Canada during the introduction of the 13-valent pneumococcal conjugate vaccine, 2010. Can J Microbiol 2012;58:1008–17. [7] Tomczyk S, Bennett NM, Stoecker C, et al. Use of 13-valent pneumococcal conjugate vaccine and 23-valent pneumococcal polysaccharide vaccine among adults aged >/=65 years: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep 2014;63:822–5. [8] Pelton SI, Weycker D, Farkouh RA, Strutton DR, Shea KM, Edelsberg J. Risk of pneumococcal disease in children with chronic medical conditions in the era of pneumococcal conjugate vaccine. Clin Infect Dis 2014;59:615–23. [9] Pilishvili T, Zell ER, Farley MM, et al. Risk factors for invasive pneumococcal disease in children in the era of conjugate vaccine use. Pediatrics 2010;126: e9–e17. [10] Shea KM, Edelsberg J, Weycker D, Farkouh RA, Strutton DR, Pelton SI. Rates of pneumococcal disease in adults with chronic medical conditions. Open Forum Infect Dis. 2014;1. ofu024. [11] Simberkoff MS, Cross AP, Al-Ibrahim M, et al. Efficacy of pneumococcal vaccine in high-risk patients. Results of a veterans administration cooperative study. N Engl J Med 1986;315:1318–27. [12] Curcio D, Cane A, Isturiz R. Redefining risk categories for pneumococcal disease in adults: critical analysis of the evidence. Int J Infect Dis 2015;37:30–5. [13] Robinson KA, Baughman W, Rothrock G, et al. Epidemiology of invasive Streptococcus pneumoniae infections in the United States, 1995–1998: opportunities for prevention in the conjugate vaccine era. JAMA 2001;285:1729–35. [14] CDC. Health, United States, 2014. Vol. ; 2015. [15] Morrill HJ, Caffrey AR, Noh E, LaPlante KL. Epidemiology of pneumococcal disease in a national cohort of older adults. Infect Dis Ther 2014;3:19–33. [16] Williams WW, Lu PJ, O’Halloran A, et al. Noninfluenza vaccination coverage among adults – United States, 2012. MMWR Morb Mortal Wkly Rep 2014;63:95–102. [17] Flory JH, Joffe M, Fishman NO, Edelstein PH, Metlay JP. Socioeconomic risk factors for bacteraemic pneumococcal pneumonia in adults. Epidemiol Infect 2009;137:717–26.

As observed in this study, PPSV23 appeared to increase, rather than decrease, IPD risk. This phenomenon likely results from the combination of residual confounding, and a vaccine that is not highly effective for IPD prevention. Per ACIP recommendations, PPSV23 was given to such a large proportion of those at highest risk for IPD (86% of the IPD cohort cases over 65 years of age in our study) that it is not possible to estimate PPSV23 effectiveness in this model. There were too few cases to calculate IPD RRs for cochlear implants or sickle cell. In general, the specific RRs for each condition, rather than the stacking model, are more accurate reflections of actual risk. While hopefully providing a clinicallyuseful tool for risk stratification, our stacking model does not provide the same detail for each condition examined, and is therefore not as accurate for each condition. We did not adjust for yearly variation in the burden of influenza-like illness. Finally, our model did not account for other chronic conditions, or socioeconomic level; these parameters may be important in estimating IPD risk [17]. 5. Conclusions This study provides IPD RRs for underlying chronic medical conditions, while adjusting for other variables related to IPD. In so doing, it validates a clinically-useful tool for assessing the RR of IPD in persons with P1chronic medical conditions. The observed RR for any of the IC medical conditions used in our analysis was >2. Risk-stacking analysis revealed that persons with >1 one nonIC medical condition carried a substantial increase in IPD RR, RRs of approximately 3 with 2 conditions, and 5 with 3 conditions. Independent of chronic conditions, one-half of all IPD cases occurred in those aged P65 years, despite a high-level PPSV23 vaccination coverage, while individuals aged 50 to <65 with non-IC medical conditions demonstrated a RR twice that of their healthy counterparts. The ACIP currently recommends a first dose of PCV13 beginning at age 65 for all persons, and before age 65 for those with IC conditions. However, those 665 years of age, with non-IC conditions remain the only age-specific and medical condition-diagnosed

Funding This work was supported by a research grant from Pfizer Inc. Acknowledgments Potential conflicts of interest: RB and NPK report research grants from Sanofi-Pasteur, Pfizer, Merck & Co., Novartis, Protein Science, Nuron Biotech, and MedImmune for other vaccine research. VS and BA are employees of Pfizer Inc and may hold stock. AY, LA, and BF have no conflicts to report. References

Please cite this article in press as: Baxter R et al. Risk of underlying chronic medical conditions for invasive pneumococcal disease in adults. Vaccine (2016), http://dx.doi.org/10.1016/j.vaccine.2016.07.003