Use of Influenza and Pneumococcal Vaccines in People with Diabetes

Reviews/Commentaries/Position Statements TECHNICAL REVIEW

Use of Influenza and Pneumococcal Vaccines in People With Diabetes

STEVEN A. SMITH, MD ef ficacy of vaccination. We will then sys- GREGORY A. POLAND, MD tematically review the clinical rep o r ts of infection, vaccine immunogenicity, and vaccine effectiveness for influenza and pneumococcus in people with diabetes. F i n a l l y, we will review general vaccine eo n a r d Thompson, the first individual pr oviders are not convinced of the threat of implementation strategies. to receive insulin for diabetes, died in ce r tain infectious illnesses or the likelihood Because most studies rep o r ting patient L1935 from complications related to a of benefit from vaccination. Alterna t i v e l y , co h o r ts and outcomes give limited and vari- staphylococcal pneumonia that occurred immunization of individuals with diabetes able information about patients with dia- after an episode of presumed influenza may be one of the many clinical proc e s s e s betes, we felt that combining res e a r ch res u l t s (1,2). Before the clinical use of insulin and of care that has been shown to be deficient (a meta-analysis) would not be valid. In antibiotics, infection often pre c i p i t a t e d for this group of patients (7–10). addition to our concerns about this hetero- ketoacidosis and other acute metabolic Initially for practical reasons and later ge n e i t y , confounders such as age and under- complications and was responsible for sig- for ethical reasons, there have been few lying cardi o p u l m o n a r y disease make risk nificant morbidity and mortality in people p l a c e b o - c o n t rolled studies designed to assessment of diabetes by itself diffi c u l t . with diabetes (3). With the introduction of demonstrate the efficacy of influenza and Th e re f o r e, we have elected to provide a nar- insulin, oral hypoglycemics, and antibi- pneumococcal immunization specifically in rative rev i e w . Primary articles were identified otics, the end-stage complications of eye, individuals with diabetes. In addition, there fr om a Medline literature search without ne u r ologic, kidney, and vascular disease ar e few clinical studies that have examined time restriction, using the key words immu- have become the major cause of death and the benefit of glycemic control, antibiotics, nization, vaccination, influenza, pneumo- su f fering for people with diabetes. In addi- antiviral agents, and other nonvaccine mea- coccus, and diabetes. Additional source s tion to the introduction of effective thera- su r es in preventing and controlling infec- we r e then identified from the ref e re n c e s pies for metabolic control and infectious tion. Current vaccine recommendations for listed from these primary sources. Search e s disease, the Advisory Committee on Immu- people with diabetes are based on historical we r e perfo r med using the last 2 years of nization Practices (ACIP) has re c o m- ob s e r vations in studies from population Cu r rent Contents. Further information was mended that individuals with diabetes gr oups that have included only a minority also obtained from contacts with the pri- receive influenza and pneumococcal of patients with diabetes. What is the evi- ma r y authors of published arti c l e s . immunization because they are reg a r ded as dence that people with diabetes benefit a moderate- to high-risk population grou p . fr om these immunizations? DIABETES AND RISKS OF In t e rn a t i o n a l l y , there are diffe re n c e s For the purpose of this rev i e w , we will I N F E C T I O N — The primary public reg a r ding vaccine recommendations for examine the impact of two vaccine-pre- health goal of immunization is to prev e n t people with diabetes, and few countries ventable illnesses: influenza and pneumo- morbidity and mortality as well as trans- have vaccine policies endorsed by medical coccal infection. We will first examine the mission of infectious disease within a pop- or patient organizations (4). Where such evidence reg a r ding the immune res p o n s e ulation. It has been rep o r ted for many years policy statements endorsing influenza and to viral and bacterial illness in people with that people with diabetes suffer significant pneumococcal immunization for moder- diabetes. We will also review the evidence morbidity and mortality from bacterial and ate- and high-risk individuals exist, many for the risks of infection in diabetes as it viral infections (11,12). This risk has been countries (including the U.S.) have defi- relates to antibody response, cell-mediated attributed to genetic and metabolic abnor- ciencies in vaccine distribution and/or im m u n i t y , leukocyte function, colonization malities—in parti c u l a r , poor glycemic con- immunization rates (4–6). This fact sug- rates, epidemiologic evidence for infection, tr ol and acidemia (13–16). Other factors gests that patients and/or their primary immune response to immunization, and associated with diabetes (age, renal disease, and cardiovascular disease) have been Fr om the Departments of Internal Medicine (S.A.S., G.A.P.), Clinical Pharma c o l o g y — V accine Research shown to be significant comorbid factors (G . A . P .), and Health Services Evaluation (S.A.S), and the Divisions of General Internal Medicine (G.A.P.) and that can increase the risk of sequelae of cer- En d o c r i n o l o g y , Metabolism, and Nutrition (S.A.S.), Mayo Clinic Rochester, Rochester, Minnesota. tain illnesses. Because of these factors, it is Ad d r ess correspondence and reprint requests to Steven A. Smith, MD, Mayo Clinic, 200 1st St. SW, W18, di f ficult to exclude the potential impact of Ro c h e s t e r , MN 55905. E-mail: [email protected]. This paper was peer-r eviewed, modified, and approved by the Professional Practice Committee, October 1999. these factors when assessing diabetes and S.A.S. acts as principal investigator for a clinical protocol supported by the Mayo Clinic and Novo Nordi s k the risks for viral and bacterial disease. Copenhagen. G.A.P. has received grant support from and has acted on an advisory panel for Merck. G.A.P. It has been assumed that specific aber- chairs the Coalition for Adult Immunization, which receives unrestricted educational funds from Merck . rations in host defense mechanisms (anti- Ab b re v i a t i o n s : AC I P , Advisory Committee on Immunization Practices; CQI, continuous quality improv e - ment; DKA, diabetic ketoacidosis. body response, cell-mediated immunity, A table elsewhere in this issue shows conventional and Système International (SI) units and conversion leukocyte function, and colonization rates) factors for many substances. account for the increased case fatality rate

DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 95 Technical Review that results from bacterial and viral infec- lococcal pneumonia (1), it is difficult to and other administrative identifiers (which tions such as influenza and pneumococcus. assess colonization’s specific impact on the ar e not validated in most studies) (39). This Poor antibody response has been shown morbidity and mortality of disease for peo- is particularly problematic with reg a r d to to be a marker of immunologic susceptibil- ple with diabetes. influenza, where viral cultures are rarel y ity to infection and its sequelae (17). Immu- done or rep o r ted. Ver y few studies have nity (as measured by in vitro opsonic activity Epidemiological evidence included serological or viral culture confir- of serum from vaccinated patients), how- Th e r e are limited studies rep o r ting the inci- mation of infection, and in those studies ev e r , has not always been correlated with dence, morbidity, and mortality of viral and that have, sufficient clinical detail to identify level of antibody measured by radioim- bacterial infections specifically in patients and assess outcomes in patients with dia- munoassay (18,19). Antibody responses to with diabetes. Tables 1 and 2 summarize the betes is often not included (40,41). Wit h the pneumococcal vaccine in people with ob s e r vational studies that have rep o rt e d reg a r d to pneumococcal infection, identifi- diabetes have been examined and have been morbidity and mortality statistics of popula- cation of these infections is most often based shown to be no diffe r ent than in people tion groups for influenza and pneumococcal on the identification of the organism (and without diabetes (20–22). Observations of infections. Because most of these studies se r ological typing) from sterile body fluids im p a i r ed antibody response to the influenza rep o r t incidence rates for groups of patients (blood, cereb r ospinal, or pleural) and, thus, vaccine, in particular when there is poor with comorbid conditions not limited to dia- a larger prop o r tion of patients who did not metabolic control, has led some to question betes (classified as high-risk, moderate-risk, have cultures perfo r med or who had infec- the value of the influenza vaccination in low-risk, immunocompetent, or immuno- tion limited to the lung or bronchial tree people with diabetes (23–26). co m p ro m i s e d ) and infrequently by specific could have been missed. This systematic Cell-mediated abnormalities (and diseases, there are limited morbidity and bias again could overestimate diabetes as a related humoral responses), such as m o rtality data specifically concern i n g risk factor. de c r eased CD4/CD8 lymphocyte ratios, patients with diabetes. A number of cau- Additional concerns for all studies changes in natural killer cell function, ti o n a r y comments reg a r ding the quality of rep o r ted include the number of patients reduced lymphocyte blastogenesis, acquired data and evidence included in each study are investigated in some population groups, the defects in interleukin-2 production, and a pr ovided in Tables 1–4. Incidence rates and techniques for sampling, the details reg a r d- reduced phagocytic function of monocytes, responses to immunization for influenza and ing estimates of excess morta l i t y , and the have all been rep o r ted in people with dia- pneumococcal disease (when available) in identification of a control population (if one betes (27–30) and could account for an patients with diabetes are rep o r ted in paren - was included). In many other series, in c r eased risk of infection (both bacterial theses. These significant limitations make although there are ref e r ences to diabetes in and viral) and poor antibody response to in t e r p r etation of available evidence prob - case fatality rates, there are insufficient details vaccination. It has also been proposed that lematic. There are no rep o r ts reg a r ding the to make more than general conclusions (41). cell-mediated immunity may be a more rel i - risks from these infections and their specific able predictor of adequate immune res p o n s e relation to glycemic control . DIABETES AND as it relates to viral infections such as Ep i d e m i o l o g i c a l l y , it has been observe d IMMUNE RESPONSE TO influenza infection (25). that influenza epidemics are associated with VA C C I N AT I O N — Th e r e have been Leukocyte function is important in the an increased number of hospitalizations and di ff e r ences in the rep o r ted immune res p o n s e s pr i m a r y immune response to bacterial infec- deaths from a variety of conditions, includ- to vaccines in people with diabetes. Patients tions (both primary and secondary infec- ing diabetes. There are case series describing with diabetes have been rep o r ted to have a tions complicating viral infections). Poorly an increased incidence of diabetic ketoaci- “slow” response to immunization (11), co n t r olled diabetes has been associated with dosis (DKA) during epidemic years (37,38). defined as the delayed but similar vaccina- al t e r ed granulocyte function (31–33). It Most of the studies rep o r ting on influenza tion-induced immune titers in patients with appears from these studies that the complex (T able 1) and all of those cited for pneumo- diabetes as compared with healthy control cellular processes of bacteria engulfment and coccal disease (Table 2) are hospital-based subjects. Because of a variety of T-cell abnor- killing could be influenced by glycemic con- co h o r ts. Because diabetes is a frequent cause malities (with and without good metabolic tr ol independent of active infection. Bacter- for hospitalization, this potential bias could co n t r ol), it has been rep o r ted that some ial colonization, if significant, could ov e r estimate the incidence of diabetes in these patients may have impaired capacity to pro- contribute to patients’ risks for complicating types of infections and lead one to suspect duce circulating B-cells and specific IgM and nosocomial infection from primary infec- that diabetes is a risk factor for susceptibility. IgG antibody in response to a vaccine tions. Patients with diabetes have been Although this type of bias in observa t i o n a l (27–30,42,43). Limited data suggest that a shown to have a higher incidence of positive studies may not allow one to adequately su b g r oup of patients who do not respond to su r veillance cultures (and carrier rates) for answer questions about diabetes, immune initial immunization may respond with coagulase-positive St a p h y l o c o c c u s and Grou p function, and risk of disease, such observa - repeated vaccination (43). It has also been B streptococci (34), whereas others have tions are helpful in planning population- stated that because the immune response to rep o r ted a colonization rate for people with based immunization strategies and in pneumococcal polysaccharide vaccines is diabetes that is no higher than that expected ta r geting people at risk for hospitalization T-cell independent, it may be normal in peo- for other individuals (35). Carrier status because of these infections (see VAC C I N E ple with diabetes (20). appears to be associated with glycemic con- IM P L E M E N T ATI O N ST R AT E G I E S ). It is assumed that an adequate immune tr ol (34,36). While it is tempting to specu- Another inherent problem with many response is necessary to benefit from vacci- late how colonization may have contributed clinical studies is the nature of the estimates nation. This point has not always been to Leonard Thompson’s complicating st a p h y - based on case identification by billing codes ap p r eciated by some who have argued that

96 DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 Smith and Poland

Table 1—Influenza and diabetes: epidemiological observations of morbidity and mortality

Yea r s St u d y Po p u l a t i o n Sample size Pr i m a r y outcome Re s u l t s Author (ref e re n c e ) ob s e rv e d ty p e * st u d i e d Co u n t r y (% diabetes) me a s u re s (d i a b e t e s ) Co m m e n t s † Stocks and Camb (52) 19 2 1 – 1 9 3 1 Po p B An n u a l En g l a n d — De a t h A, B, O Re g i s t e r Wal e s Excess morta l i t y (6–1 4 % ) Finland et al. (137) 19 4 0 – 1 9 4 1 Ca s e S Boston City U. S . 66 (3) Staph pneumonia A, C, O Ho s p i t a l St u a rt - H a r ris et al. (89) 19 4 9 – 1 9 5 0 Ca s e S Hospital and En g l a n d 85 De a t h — C, E, F, O ot h e r Pre v a l e n c e (4 . 5 % ) Ei c k h o f f et al. (53) 19 5 5 – 1 9 6 0 Po p B Re g i s t r a r U. S . — De a t h A National Offi c e Excess morta l i t y (4 – 1 4 % ) of Vital Statistics Ho u s w o r th and 19 5 7 – 1 9 6 6 Po p B Re g i s t r a r U. S . — De a t h A Langmuir (55) National Offi c e Relative intensity (0 . 5 – 7 . 7 ) of Vital Statistics Ma r tin et al. (138,139) 19 5 7 – 1 9 5 8 Ca s e S Hospitals in U. S . 32 (6.25) De a t h — C, J, O Bo s t o n Pe t e r s d o r f et al. (54) 19 5 7 Ca s e S New Haven U. S . 91 (4.4) Pn e u m o n i a A, C, O Ho s p i t a l de a t h Case fatality 12% (0%) Oseasohn et al. (140) 19 5 7 Ca s e S Autopsy series, U. S . 33 (3) De a t h — A, E, O Cl e v e l a n d ho s p i t a l s Giles and 19 5 7 Ca s e S Autopsy series En g l a n d 46 (4.3) De a t h — A, C, H, O Sh u t t l e w o r th (64) No rt h St a ff o rd s h i r e Barker and 19 6 8 – 1 9 7 3 CC HM O , U. S . 38 (13) Pn e u m o n i a — A, K, I, O Mullooly (59) Po r tland, Oreg o n De a t h Sc h w a r zmann 19 6 8 – 1 9 6 9 CC Grady Memorial U. S . 108 (14) Pn e u m o n i a — A, C, O et al. (40) Hospital, Atlanta Bisno et al. (41) 19 6 8 – 1 9 6 9 Ca s e S City of Memphis U. S . 79 (10) Pn e u m o n i a C, D, G, J, O ho s p i t a l s Case fatality 13 % Ca m e r on et al. (56) 19 6 9 – 1 9 8 1 CC Australia Burea u Au s t r a l i a 4, 0 9 5 De a t h A, G of Statistics Odds ratio (2 . 0 ) 95% CI (0 . 4 – 1 4 . 8 ) Watkins et al. (37) 19 6 9 – 1 9 7 0 Ca s e S Ho s p i t a l s , En g l a n d 29 (100) DK A C, E, O Bi rm i n g h a m De a t h Case fatality (2 4 % ) Bouter et al. (38) 19 7 6–1 9 7 9 CC Dutch National Ne t h e r l a n d s — DK A A, F Medical Registrar Relative risk (1 5 . 9 ) Pn e u m o n i a Relative risk (2 5 . 6 ) Ho s p i t a l i z a t i o n 95% CI (5 . 7 – 6 . 2 ) De a t h 95% CI (4 2 . 4 – 9 1 . 8 ) Ca r rat and 19 8 6 – 1 9 9 0 Po p B National Center Fr a n c e 7, 7 0 0 De a t h A, E Val l e r on (58) of Mortality Ra t e s 48 1 . 5 – 3 7 8 . 8 St a t i s t i c s (6 . 2 – 5 . 6 ) / 1 0 5 McBean et al. (57) 19 8 9 – 1 9 9 1 CC Me d i c a r e U. S . 30 10 5 Influenza A A, L hospital billing Ho s p i t a l i z a t i o n rec o rd s Relative risk (1 . 0 4 ) 95% CI (1 . 0 – 1 . 0 8 ) De a t h Relative risk (1 . 6 4 ) 95% CI (1 . 2 1 – 2 . 2 1 ) continued on page 98

DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 97 Technical Review

Table 1 (c o n t i n u e d )

Yea r s St u d y Po p u l a t i o n Sample size Pr i m a r y outcome Re s u l t s Author (ref e re n c e ) ob s e rv e d ty p e * st u d i e d Co u n t r y (% diabetes) me a s u re s (d i a b e t e s ) Co m m e n t s † McBean et al. (57) 19 8 9 – 1 9 9 1 CC Me d i c a r e U. S . 30 10 5 Influenza B A, L Hospital billing Ho s p i t a l i z a t i o n rec o rd s Relative risk (0 . 9 8 ) 95% CI (0 . 9 5 – 1 . 2 0 ) De a t h Relative risk (1 . 5 1 ) 95% CI (1 . 1 3 – 2 . 0 2 ) Diepersloot et al. (2) — R — —— Pn e u m o n i a —O Dea t h Neuzil et al. (141) 19 7 3 – 1 9 9 3 Po p B No n p re g n a n t U. S . 21 10 6 Ca rd i o p u l m o n a r y A, B, F, L women 15–64 (6 . 5 % ) Ev e n t s years of age Relative risk (1 6 . 4 ) Ten n e s s e e 95% CI (1 1 . 9 – 2 0 . 9 ) Me d i c a i d *CC, case control; CaseS, case series; PopB, population-based descriptive; R, rev i e w . †A, no or incomplete serological confirmation of infection, based on clinical diagnosis; B, methodology of calculating expected mortality (use of “annual decrements or increments thus calculated from the decade . . . assumed to continue constant”); C, ref e r ral: hospital-based cohort; D, historical control subjects or from a diffe r ent population group; E, insufficient rep o r ting of details to confirm details concerni n g patients with diabetes; F, selection method: selection method–rep o r ting of a convenience rep r esentative sample from cohort; G, selection method: nonrandom or without mention of selection technique for sample cohort; H, selection method: one of two patients with diabetes had acromegaly and suspect secondary diabetes; I, selection method: HMO population; J, selection method: results section indicated that case finding was incomplete; K, incomplete medical rec o r d review: results section indicated missing medical rec o r ds; L, incomplete medical rec o r d review: administrative data set only without medical rec o r d validation; O, depended exclusively or heavily on narrative rev i e w . HMO, health maintenance orga n i z a t i o n . because most people with diabetes have immunization (44,45–48). With this method, viduals who have more health concerns ap p r opriate immune responses, they do not immunization efficacy is determined by (and healthy lifestyles) may also actively seek have an indication for immunization (17). comparing the distribution of pneumococcal and elect to receive immunization. Finally, On the contrary, if people with diabetes have se r otypes that cause infection in vaccinated many studies use recall as a method of doc- ap p r opriate responses to vaccination and and unvaccinated individuals (assuming that umentation of vaccination, despite the fact evidence for increased risk from vaccine-pre- the risk of infection with vaccine-type orga n - that the accuracy of patient historical rec a l l ventable illness such as influenza and pneu- isms is similar in both populations). has been estimated to be 80% at best (50). mococcal disease, it is reasonable to assume Because of the lack of randomized clini- that immunization may be of value. cal trials, there is inherent bias in these types IN F L U E N Z A of studies of vaccine effi c a c y . Individuals cho- DIABETES AND EFFICACY OF sen as control subjects (in case-control series) Clinical reports of infection VAC C I N AT I O N — It is unlikely that a ar e likely to have had circumstances that The first descriptions of large outbreaks of randomized placebo-controlled study of lead to their immunization (or lack of influenza were re p o rted in 1847–1848 influenza and pneumococcal immunization immunization). Individuals receiving vac- (51). The death rate due to influenza epi- in people with diabetes will be perfo r med in cine are more likely to have comorbid con- demics remained high during the following the U.S. because of the large number of ditions such as diabetes, whereas others may decade and then steadily declined in the patients req u i r ed and ethical questions of have not received vaccination and have had subsequent 30 years until influenza as a randomization to placebo. Because of this worse outcomes because of other character- re c o rded cause of death almost disap- fact, studies to answer the question of istics that are obstacles to timely health care pe a r ed. With the epidemic of 1890–1891, immunization efficacy have fallen under and immunization (such as dementia or an increase in death rate and significant th r ee categories (Tables 3 and 4): clinical tri- st r oke) (49). Additional limitations for many morbidity associated with influenza epi- als involving at-risk patient groups (which studies include the bias that not all people demics was again observed (51). include patients with diabetes), case-con- with diabetes are rep r esented in studies of Until the 20th century, there was lim- tr ol studies, and indirect cohort analysis institutionalized patients (e.g., nursing home ited information re g a rding the risks of (44,45). Although case-control studies pro- or hospital). Diffe r ences in opportunities for influenza in people with diabetes. Stocks vide a practical means for achieving ade- ex p o s u r e to contagious illnesses such as and Camb (52) were two of the first inves- quate statistical power, they often include a influenza may bias some rep o r ts. Selection tigators in the 1900s (1921–1931) to rep o r t he t e r ogeneous group of at-risk patients and bias exists for cohort studies from ref e rr a l an increase in death rate (6–14%) from thus provide limited information about vac- hospitals because these patients are more influenza in people with diabetes. Eickhoff cine efficacy specific for people with dia- often severely ill (and thus more likely to be et al. (53) rep o r ted 20 years later that peo- betes. The indirect cohort method has been hospitalized) and in turn may be more likely ple with diabetes had a moderate increase in fr equently used in studying pneumococcal to be immunized. In contrast, other indi- death during a 6-month influenza outbrea k

98 DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 Smith and Poland

Table 2—Pneumococcal infection and diabetes: epidemiological observations of morbidity and mortality

Yea r s St u d y Po p u l a t i o n Sample size Pr i m a r y outcome Re s u l t s Author (ref e re n c e ) ob s e rv e d ty p e * st u d i e d Co u n t r y (% diabetes) me a s u re s (d i a b e t e s ) Co m m e n t s † Austrian and Gold (15) 19 5 2 – 1 9 6 2 Ca s e S Kings County U. S . 457 (4.6) Pn e u m o n i a — A, C, F, O Ho s p i t a l , Sep t i c e m i a Bro o k l y n Dea t h Mufson et al. (101) 19 6 7 – 1 9 7 0 Ca s e S Cook County U. S . 325 (3.7) Pn e u m o n i a — A, C, O Hospital, Chicago Extra pulmonary infection Sep t i c e m i a Dea t h Case fatality rate 28% (42%) Fedson and 19 7 0 – 1 9 8 1 Ca s e S University of U. S . 205 (12.3) Pn e u m o n i a — A, C, K, O Ch i a r ello (91) Ch i c a g o Sep t i c e m i a ho s p i t a l s , Dea t h Ch i c a g o Case fatality rate 21% (32%) Gransden et al. (103) 19 7 0 – 1 9 8 4 Ca s e S St. Thomas En g l a n d 325 (3.3) Se p t i c e m i a C, O Ho s p i t a l Case fatality rate 29% (9.1%) Finkelstein et al. (92) 19 7 2 – 1 9 8 1 Ca s e S Bellevue Hospital U. S . 187 (9.1) Se p t i c e m i a — C, G, K, O Hook et al. (102) 19 7 4 – 1 9 8 0 Ca s e S Harbor Vie w U. S . 134 (4) Se p t i c e m i a C, K, O Medical Center, Dea t h Se a t t l e Case fatality rate 31% (40%) Gr uer et al. (104) 19 7 4 – 1 9 8 1 Ca s e S E. Birmi n g h a m En g l a n d 103 (2) Se p t i c e m i a — C, K, O Ho s p i t a l Mylotte and Beam (93) 19 7 7 – 1 9 8 0 Ca s e S Veterans Hospital, U. S . 62 (19) Pn e u m o n i a — C, H, O Bu ff a l o Sep t i c e m i a Dea t h Kramer et al. (14) 19 7 7 – 1 9 8 5 Ca s e S Sh a a r e Zedek — 55 (1) Pn e u m o n i a — C, H, O and R Medical Center Sep t i c e m i a Dea t h Case fatality rate 36% (67%) Wat a n a k u n a k o r n 19 8 0 – 1 9 8 9 Ca s e S Co m m u n i t y U. S . 385 (12) Pn e u m o n i a — C, G, O et al. (94) and R ho s p i t a l ( s ) , Sep t i c e m i a Oh i o Dea t h Case fatality rate 25% (37%) Odds ratio 1. 0 ‡ SE 0. 4 Al v a r ez et al. (95) 19 8 0 – 1 9 8 5 Ca s e S Veterans Hospital, U. S . 55 (12.7) Se p t i c e m i a — C, H, O Johnson City Perlino and 19 8 0 – 1 9 8 1 Ca s e S Grady Memorial U. S . 93 (10.8) Pn e u m o n i a — C, H, O Rimland (16) and Vet e r a n s Sep t i c e m i a Hospital, Atlanta Dea t h *CaseS, case series; R, rev i e w . †A, no or incomplete strain typing for confirmation of infection, based on clinical diagnosis; C, ref e r ral: hospital-based cohort; F, selection method: cohort with positive blood cultures orde r ed only per physician/clinical suspicion; G, selection method: nonrandom or without mention of selection technique for sample cohort; H, selection method: small cohort; K, incomplete medical rec o r d review: results section indicated missing medical rec o r ds; O, depended exclusively or heavily on narrative rev i e w . ‡P = 0.05.

(1957–1958), while Petersdorf et al. (54) et al. (38) rep o r ted two times the number of during both epidemic and nonepidemic rep o r ted a series of 91 patients (4.4% with patients with diabetes dying after hospital- co n t r ol periods (40,54,58). The average hos- diabetes) who had a 12% incidence of death ization for pneumonia or DKA in the pitalization rate from 1989 to 1991 was 4 fr om pneumonia (but none of whom had influenza epidemic years of 1976 and 1978 per 1,000 Medicare beneficiaries who had diabetes). In addition, Housworth and as compared with the nonepidemic years of the diagnosis of diabetes (57). Langmuir (55) found that excess deaths in 1977 and 1979. In other series of hospital- Although it is rep o r ted that patients individuals with diabetes were significant ized patients, the incidence of diabetes in with diabetes are at increased risk of death (i n c r eased by 5–12%) in six of seven epi- influenza epidemic and nonepidemic years fr om influenza and complicating pneumo- demics from 1957 to 1966 (excluding has been remarkably constant (54,57). nia (38,52,53,55,56), in several series, age 1966). Cameron et al. (56) rep o r ted the These rep o r ts suggest that people with dia- and cardiac and pulmonary disease have odds ratio for death in patients with diabetes (in comparison to individuals without been the greatest risks for death (7,41). betes as 2.0 (95% CI 0.4–14.8), and Bo u t e r similar chronic disease) are at increased risk Ch r onic obstructive pulmonary disease and

DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 99 Technical Review

Table 3—Influenza and diabetes: efficacy of immunization

Yea r s St u d y Po p u l a t i o n Sample size Pr i m a r y outcome Re s u l t s Author (ref e re n c e ) ob s e rv e d ty p e * st u d i e d Co u n t r y (% diabetes) me a s u re s (d i a b e t e s ) Co m m e n t s † Barker and 19 6 8 – 1 9 7 3 RC HMO, Californi a U. S . 1, 1 0 0 Ho s p i t a l i z a t i o n Vac c i n e A, E, I Mullooly (81) (1 2 – 1 5 ) Yes No 19 6 8–1 9 6 9 Ra t e s 4/ 1 , 1 2 0 15 / 2 , 8 2 0 19 7 2 – 1 9 7 3 Ra t e s 2/ 1 , 1 0 0 24 / 3 , 7 0 0 § De a t h 19 6 8 – 1 9 6 9 Ra t e s 3/ 1 , 1 2 0 1/ 2 , 8 2 0 19 7 2 – 1 9 7 3 Ra t e s 0/ 1 , 1 0 0 3/ 3 , 7 0 0 § Saah et al. (82) 19 7 9 – 1 9 8 2 RC Nursing home, U. S . 1, 3 6 2 Pn e u m o n i a A, C, E, G, O Ma n h a t t a n (1 4 . 5 ) De a t h No vaccine 19 7 9 60-Day risk ratio 3. 3 § 19 8 0 60-Day risk ratio 2. 9 § 19 8 1 60-Day risk ratio 0. 9 Mullooly et al. (83) 19 8 0 – 1 9 8 9 RC HMO, Californi a U. S . 25 1 , 0 3 4 Pn e u m o n i a No vaccine A, H, E, I, L pe r s o n - p e r i o d s Odds ratio 1. 4 95% CI 1. 2 – 1 . 7 3 De a t h Odds ratio 1. 4 9 95% CI 0. 9 – 2 . 3 6 Fedson et al. (73) 19 8 2 CC No n i n s t i t u t i o n a l , Ca n a d a 976 (5.8) Pn e u m o n i a No vaccine A, D, E, J, L Ma n i t o b a Relative risk 1. 4 7 95% CI 1. 0 9 – 2 . 0 8 Influenza death Relative risk 2. 8 6 95% CI 1. 2 8 –6. 2 5 19 8 5 878 (5.9) Pn e u m o n i a Relative risk 1. 6 1 95% CI 1. 2 3 – 2 . 0 8 Influenza death Relative risk 2. 1 7 95% CI 1. 0 8–4 . 5 5 Ng u y e n - Va n - T am 19 8 9 – 1 9 9 0 CC Le i c e s t e r s h i r e En g l a n d 15 6 Ho s p i t a l i z a t i o n No vaccine A, C, D, K et al. (84) Ho s p i t a l Odds ratio (1 . 4 8 ) 95% CI (0 . 7 – 3 . 3 4 ) Foster et al. (74) 19 8 9 – 1 9 9 0 CC Me d i c a r e vaccine U. S . 721 (13) Ho s p i t a l i z a t i o n No vaccine A, B, K, L demo, Michigan Odds ratio (1 . 2 ) 95% CI (0 . 8–1 . 8 8 ) Monto et al. (85)‡ 19 8 9 – 1 9 9 0 CC Me d i c a r e vaccine U. S . 721 (13) Ho s p i t a l i z a t i o n No vaccine A, B, K, L demo, Michigan “height of ep i d e m i c ” Odds ratio (0 . 9 7 ) Nicholson et al. (87) 19 8 9 – 1 9 9 4 CC Le i c e s t e r s h i r e En g l a n d 80 (100) Ho s p i t a l i z a t i o n Vac c i n e A, D, G, J, L hospital reg i s t e r Odds ratio (0 . 2 1 ) 95% CI (0 . 0 5 –0. 8 1 ) Strikas et al. (50) 19 9 0 – 1 9 9 1 CC No n i n s t i t u t i o n a l U. S . 48 1 Pn e u m o n i a A, B, C, G Me d i c a r e billing community control rec o r ds, Ohio su b j e c t s No vaccine and Pennsylvania Odds ratio (1 . 0 7 ) Hospital control su b j e c t s Odds ratios (0 . 8 ) Nichol et al. (49) 19 9 0 – 1 9 9 1 CC Gr oup health U. S . 25 , 0 0 0 Ho s p i t a l i z a t i o n A, E, I, H 19 9 1 – 1 9 9 2 HMO, St. Paul per cohort Pne u m o n i a 19 9 2 – 1 9 9 3 Total: 78,527 Hea r t failure “costs” (6 . 4 – 1 1 . 6 ) Dea t h Vac c i n e Rate red u c t i o n 27 – 5 7 % *CC, case control; RC, ret r ospective cohort. †A, no or incomplete serological confirmation of infection, based on clinical diagnosis; B, methodology: use of survey and self-rep o r t- ing for diagnoses/vaccination; C, ref e r ral: hospital nursing home–based cohort; D, bias: control group; E, insufficient rep o r ting of details to confirm details concerning patients with diabetes: odds ratio and relative risk expressed for “high risk” conditions and risk ratio for entire group; F, selection method: selection method–rep o r ting of a convenience rep r esentative sample from cohort; G, selection method: nonrandom or without mention of selection technique for sample cohort; H, selection method: high incidence of other comorbid conditions (pulmonary and cardiac disease); I, selection method: HMO population; J, selection method: results section indicated case finding was incomplete; K, incomplete medical rec o r d review: results section indicated missing medical rec o r ds; L, incomplete medical rec o r d review: administrative data set only without medical rec o r d validation; O, depended exclusively or heavily on narrative rev i e w . ‡Apparent same cohort as Reference 84. §P 0.05; NS. HMO, health maintenance orga n i z a t i o n .

100 DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 Smith and Poland

ca r diovascular disease as causes for hospi- dying from influenza during hospitalization tion has been shown to be effective in pre- talization have been the most freq u e n t l y for people with diabetes has been rep o rt e d venting hospitalization with pneumonia rep o r ted disease categories during epidemic as high as 92 or an incidence of 481 per 105 and influenza in high-risk noninstitution- influenza periods (39,59,60). Several rep o rt s patients during the epidemic year 1978 alized elderly individuals (including pa t i e n t s and case series either have not identified (38,59) and as low as 5.26–6.2 per 105 with diabetes) (73,74). diabetes as an individual comorbid condi- patients in the subsequent epidemics of tion (61) or have not found diabetes among 1985–1986 and 1989–1990 (58). Again, Vaccine effectiveness their patients (62,63). In addition, Giles and the diffe r ences may be explained in part Discussion of flu vaccine efficacy is more S h u t t l e w o rth (64) re p o rted only two because of the older age of the cohort in complex than it might seem at first. In part, patients with diabetes (4.3%) in their 1978 ( 75 years) and their higher inci- it depends on what is measured, i.e., the effi - autopsy series, one of whom had acrom e g a l y dence of rep o r ted deaths from res p i r a t o r y, cacy of flu vaccine against symptoms or as an unusual secondary cause of diabetes. ca rd i o v a s c u l a r , and end-stage renal diseases. against morta l i t y . So, for example, the effi - Morbidity and mortality from influenza Details from these studies do not allow cacy of flu vaccine against symptomatic ar e often influenced by factors other than the assessment of these factors. influenza is 70–90% in healthy adults versus pa t i e n t ’ s comorbid diagnoses. Alling et al. Conclusion. Published case series and pop- 30–40% in frail elderly individuals. How- (60) have pointed out that some epidemics ulation-based cohorts are not able to answer ev e r , the efficacy of influenza vaccine against of influenza are accompanied by a grea t e r the question about the true incidence of lower res p i r a t o r y tract involvement, hospi- excess mortality than others, partly because influenza and its complications in patients talization, and death is far higher. In addi- of viral antigenic variation such as the intro- with diabetes. In addition, they are unable to tion, efficacy varies each year because the duction of a new antigenic subtype or of answer the question reg a r ding the role of influenza vaccine is essentially a new vaccine variants derived by the process of antigenic metabolic control and its influence on the each year. Thus, efficacy also depends on drift and shift. Sabin (65) observed a attack rate and severity of illness in this same how well the vaccine strains and the circu - de c r ease in the mortality from influenza and gr oup of patients. Independent of these lim- lating strains match. Finally, efficacy can pneumonia during influenza epidemics itations, many have interpreted these data to appear falsely low depending on the pres - oc c u r ring between 1971 and 1975, and no suggest that patients with diabetes and its ence of cocirculating viruses against which excess mortality was rep o r ted from the complications (who are also likely to be the influenza vaccine is not protective (cox- influenza epidemics of 1953, 1965, 1978, patients with frequent hospitalizations) are sackie virus, adenovirus, parainfluenza, etc.). and 1979 (56). During these latter years, at increased risk for the morbidity, morta l i t y , In many intervention studies (75–78), it influenza more often caused illness in chil- and health care costs resulting from influenza has been demonstrated that vaccination dr en and young adults (56), rei n f o r cing the in f e c t i o n . against influenza (during epidemic and fact that age, a significant risk factor for nonepidemic years) is associated with less fre- influenza, is likely to bias observations con- Vaccine immunogenicity quent hospitalizations for complications of ce r ning diabetes, which has a greater inci- Influenza immunization and subsequent influenza, fewer deaths during the influenza dence and risk for complications in the immunity results from stimulation of se a son, and direct savings in health care el d e r l y . Because it is not possible to control humoral and cell-mediated immune costs. Two reviews and a metanalysis sup- for these factors in studies that have been responses (66). The antibody- and cell- po r t this conclusion (79,80), but none of rep o r ted to date, there is a potential for both mediated responses (from natural exposure these rep o r ts mentions people with diabetes an over- and underestimate of diabetes as a and immunization) in patients with diabetes as a specific population group or as part of risk factor. have been comparable to those in control an at-risk grou p . Although one series noted an increa s e subjects in some rep o r ts (24,26,66–68) but Several of the studies listed in Table 3 (t h r eefold) in hospital admissions for the not in others (21,22,27,69,70). In the pres - (with rep o r ted prevalence rates of diabetes of acute metabolic complications of DKA dur- ence of end-organ disease such as urem i a 6–14.5%) describe outcomes for heterog e - ing influenza epidemics (37), there are lim- and dialysis, response to immunization in neous at-risk patient groups and thus pro- ited population-based studies that have patients with diabetes has been rep o r ted to vide insufficient detail to discern the su p p o r ted this observation. Other metabolic be very poor (71). potential benefit of immunization specifi- ab n o r malities (hypokalemia) in people with Poor immunological response to cally for people with diabetes (49,73, diabetes have also been rep o r ted during epi- influenza vaccination has been attributed 81–83). These studies demonstrate the effi - demics and are noted to occur out of pro- to impairment of antigen-specific antibod- cacy of immunization in at-risk patients for po r tion to that usually associated with DKA ies because of nonenzymatic glycosylation most but not all epidemic influenza periods. alone (37). The largest population-based of serum IgG resulting from poor glycemic In another study, subgroup analysis of study with detailed data concluded that dur- co n t r ol (21). Others have not observe d patients with diabetes demonstrated the effi - ing epidemic years, patients with diabetes that antibody production is related to cacy of influenza vaccination that could not we r e six times more likely to be hospitalized metabolic control and suggest that failure be demonstrated in other population grou p s with a diagnosis of influenza than age- and to respond may be unique to the type of (84). In contrast, odds ratios have been sex-matched control subjects (38). diabetes and susceptibility genes for type 1 rep o r ted in favor of protection in large het- Independent of diabetes, influenza has diabetes and dysregulation of the immune er ogeneous population groups (in parti c u l a r , been shown to be associated with excess response (22,69,70,72). Despite these gr oups with concomitant heart and lung mo r tality in individuals 65 years of age speculations, a sufficient immune res p o n s e disease) but without specific benefit for and in those with cardiovascular disease has been rep o r ted in 70% of patients in patients with diabetes (74,85). Incomplete (7,53,60). The estimated relative risk of many studies (2,22,24,67), and vaccina- medical re c o rd re v i e w, dependence on

DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 101 Technical Review

Table 4—Pneumococcus and diabetes: efficacy of immunization

Yea r s St u d y Po p u l a t i o n Sample size Pr i m a r y outcome Re s u l t s Author (ref e re n c e ) ob s e rv e d ty p e * st u d i e d Co u n t r y (% diabetes) me a s u re s (d i a b e t e s ) Co m m e n t s † Br oome et al. (44) 19 7 8 – 1 9 8 0 IC Samples for U. S . 35 (3) Cu l t u r e (sterile — C, G, J, M req u e s t e d body fluid) se r otyping by % Vaccine effi c a c y 60 % CD C Sh a p i r o and 19 7 8 – 1 9 8 2 CC Yale-New Haven U. S . 90 Cu l t u r e (sterile — C, E, P Clemens (112) Ho s p i t a l body fluid) % Vaccine effi c a c y 67 % 95% CI 13 – 8 7 % Bolan et al. (46) 19 7 8 – 1 9 8 4 IC CDC surve i l l a n c e U. S . 1, 8 8 7 Cu l t u r e (sterile — B of U.S. hospitals (2 ) body fluid) % Vaccine effi c a c y 55 % (9 0 % ) Butler et al. (45) 19 7 8 – 1 9 9 2 IC and CDC National U. S . 2, 8 3 7 Cu l t u r e (sterile — B, C, M CC Su rv e i l l a n c e (5 ) body fluid) St u d y % Vaccine effi c a c y 57 % (8 4 % ) 95% CI 45 – 6 6 % (5 0 – 9 5 % ) Fo r rester et al. (47) 19 7 9 – 1 9 8 5 IC and Denver Vet e r a n s U. S . 89 Ba c t e re m i a — A, B, C, D, CC Medical Center E, J, H, K, P Sims et al. (113) 19 8 0 – 1 9 8 6 CC Five hospitals in U. S . 12 2 Cu l t u r e (sterile — B, C, E, K Ph i l a d e l p h i a (1 4 . 8 ) body fluid) % Vaccine effi c a c y 70 % 95% CI 36 – 8 6 % § Si m b e r k o f f et al. (108) 19 8 1 – 1 9 8 5 RC T Vet e r a n s U. S . 2, 3 5 4 Ba c t e re m i a A, D, N, M Ad m i n i s t r a t i o n (3 0 ) Pne u m o n i a Bron c h i t i s Relative risk 1. 1 5 (0 . 9 0 ) 95% CI 0. 8 9 – 1 . 4 8 (0 . 7 – 1 . 2 4 ) Koivula et al. (114) 19 8 2 – 1 9 8 5 PC Var k a u s Fi n l a n d 4, 1 6 7 Pn e u m o n i a A, N, E, H (1 3 . 1 ) Odds ratio 1. 6 6 95% CI 0. 7 7 – 3 . 5 7 Sh a p i r o et al. (48) 19 8 3 – 1 9 9 0 Major hospitals, U. S . 1, 0 5 4 Cu l t u r e (sterile — A, C, D, N, P Co n n e c t i c u t (4 . 6 ) body fluid) PC and % Vaccine effi c a c y 56 % IC % Vaccine effi c a c y 48 % *CC, case control; PC, prospective cohort; IC, indirect cohort; RCT, randomized control trial. †A, no or incomplete serological confirmation of infection, based on clinical diagnosis; B, methodology: use of surve y , self-rep o r ting, or not stated for diagnoses/vaccination status; C, ref e r ral: hospital nursing home–based cohort; D, bias: con- tr ol group (lower incidence of bacteremia, risk stratification, and age than match cases); E, insufficient rep o r ting of details to confirm details concerning patients with diabetes: odds ratio, relative risk, vaccine efficacy expressed for “high-risk” conditions; F, selection method: selection method–rep o r ting of a convenience rep re s e n t a t i v e sample from cohort; G, selection method: nonrandom or without mention of selection technique for sample cohort of samples of serotyping; H, selection method: high incidence of other comorbid conditions (i.e., pulmonary, cardiac, liver, renal disease); I, selection method: health maintenance organization population; J, selection method: case finding was incomplete or had very small numbers of patients; K, incomplete medical rec o r d review: missing medical rec o r ds; L, incomplete medical rec o r d rev i e w : administrative data set only without medical rec o r d validation; M, large number of cohort 18 years of age; N, testing 14-valent vaccine; P, no specific data concerni n g patients with diabetes. Not included is “% vaccine efficacy” (with 95% CI) given for a subgroup of patients with chronic disease, some with diabetes. §P 0.005; NS . CDC, Centers for Disease Control . patient surveys for documentation of both incidence of the comorbid illnesses such as Strikas et al. (50) could not document a pro- diabetes diagnosis and vaccination, and ch r onic res p i r a t o r y conditions and conges- tective effect of influenza vaccination for assigning individuals who did not ret u r n tive heart failure contributing to adverse out- high-risk patients (including diabetes) su r veys as “not vaccinated” are examples of comes, Nichol et al. (49) did not consider against influenza B during the years bias that make interpretation of the last two diabetes in their multivariate model of the 1990–1991; however, this may have been studies (74,85) difficult. Because of the high co s t - e f fectiveness of influenza vaccination. due to poorly matched vaccines and lack of

102 DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 Smith and Poland

significant disease (morbidity and morta l i t y ) and series of pneumococcal bacteremia fail Vaccine effectiveness fr om influenza during this study period. to mention individual patients with diabetes Th e r e are few studies to assess the effe c t i v e - Fi n a l l y , in many studies where vaccine effe c - (96–100). These rep o r ts either failed to clas- ness of pneumococcal vaccine in people with tiveness is described for subgroups of pop- sify their patients according to chronic dis- diabetes. Although not specific to diabetes, ulation cohorts, there is often a category of ease (96–98) or made no mention of ef ficacy for prevention of pneumococcal “metabolic diseases.” While this category diabetes among the chronic diseases listed pneumonia was shown in randomized do u - includes patients with diabetes, the addi- for their patients (99,100). ble-blind placebo-controlled clinical trials tional comorbid conditions in this category Diabetes as well as increased age, an conducted among young, otherwise healthy ( renal failure and hepatic insuff i c i e n c y ) ex t r a p u l m o n a r y site of pneumococcal infec- individuals where the disease was endemic make interpretation specifically for diabetes tion, the presence of cirrhosis, alcoholism, (e.g., South African gold miners and Air impossible (86). azotemia, and infection with certain capsular Fo r ce rec r uits) (105–107). Efficacy was not Conclusion. Definitive proof of the effi c a c y types (such as type 3) appear to contribute demonstrated in a clinical trial among U.S. of influenza vaccination specifically in peo- the most to the risk of death from bacterem i c veterans (50 years or older) with a variety of ple with diabetes is lacking. Studies that pneumococcal disease (14,91,92, ch r onic disease, including 664 patients with include diabetes as one of the at-risk patient 94,101,102). Because age has such a stron g diabetes (with relative risk of 0.8–1.92; 95% gr oups, however, support immunization in impact on the frequency of diabetes rep o rt e d CI for vaccine suptypes) (108). This is the this patient group. In parti c u l a r , these same in cohorts of bacteremic patients (102) and only randomized control trial published that studies consistently support influenza vacci- mo r tality in others (13–15,92–94,98,101– allows examination of sufficient detail con- nation when there are comorbid conditions 104), it is difficult to assess the impact of dia- ce r ning patients with diabetes. However, with diabetes such as age and cardi o v a s c u l a r betes as an independent risk factor. Case significant limitations of this study included complications. In the limited studies that fatality rates for children are typically low the use of the older 14-valent vaccine, prob - included a sufficient number of people with (94,97,100). Adults ( 50 years of age) have lems with the clinical definition of nonbac- diabetes for statistical power (87), influenza rep o r ted fatality rates of 2.4% compared with te r emic illness, infrequent bacteremic illness, immunization was effective in reducing hos- 15% in patients 50 years of age (103). and an extremely low statistical power to pital admissions during influenza epidemics. Some series have rep o r ted high fatality rates demonstrate efficacy (45,109). Because the (up to 50%) with nosocomial pneumococcal U.S. population (even those with chronic ill- PN E U M O C O C C U S ba c t e r emia (93,95), whereas others have ness) has a lower incidence of pneumococ- rep o r ted rates of 7–9% (16,94). In series cal infection than the previously studied Clinical reports of infection rep o r ting both high and low fatality rates, population groups, many have questioned Pneumococcal pneumonia is the most com- patients with diabetes have had a high inci- the efficacy of this vaccine for many chron i c mon form of acute bacterial community- dence (20–29%) of nosocomial acquired diseases (107,108,110,111). ac q u i r ed pneumonia (88). Notwithstanding infections (16,95). To overcome the problems of making a the prior discussions of influenza and dia- Conclusion. Published studies support the clinical diagnosis of nonbacteremic illness betes in this article, the pneumococcus has fact that people with diabetes are at least as and the limited statistical power associated been the organism most frequently associ- likely to be susceptible to pneumococcal with studying only infection with associated ated with pneumonia in epidemic and infection as other patients with chronic dis- ba c t e r emia, Broome et al. (44) first advo- nonepidemic influenza years (40,41,56,89). ease. Many studies suggest that diabetes may cated the use of indirect cohort analysis to Pneumococcal infection has accounted for be a unique risk factor for an increased inci- co m p a r e the serotypes of isolates from pneu- 48% of pneumonia in cases associated with dence of bacteremia associated with this mococcal infections in vaccinated and the Hong Kong influenza epidemic period org a n i s m . unvaccinated patients. Broome et al. also (1968–1969) and up to 62% in a nonepi- raised concerns about pneumococcal vac- demic control period (1967–1968) (40). Vaccine immunogenicity ci n e ’ s efficacy in individuals with chronic dis- Ba c t e r emia is seen in 8–50% of individuals In 1983, a 23-valent vaccine replaced the ease; however, in their series of 35 patients with pneumococcal infections, and of these, original 14-valent vaccine first released in (most of whom had splenectomy as an indi- 15–20% are fatal despite antibiotics (19). 1977 (45). Although antibody response to cation for vaccination), there was only one This high case fatality rate from bacterem i c pneumococcal polysaccharides may proc e e d person with diabetes (44). In addition, vac- pneumococcal disease supports the concept independent of T-cell–mediated help (22), cine efficacy would not have been expected that a reduction in the number of deaths vaccine efficacy has been questioned in this individual with diabetes who had sep- related to this infection can only be accom- because of the potential for poor metabolic sis from an underlying nonvaccine serot y p e plished by widespread immunoprop h y l a c t i c c o n t rol to influence cellular immune pneumococcal pneumonia. me a s u re s . response (delayed hypersensitivity and cyto- Additional indirect cohort and case-con- Not all agree that diabetes is a risk factor toxic T-cell response) (22,23,67). Giebink et tr ol studies (Table 4) have shown vaccine effi - for pneumococcal disease. A ret ro s p e c t i v e al. (20) rep o r ted that individuals with dia- cacy to be 77–90% for prevention of invasive su r vey of select ref e r ral clinics for people betes and complications such as nephrot i c pneumococcal infection in patients with dia- with diabetes did not support an increa s e d sy n d r ome may have appropriate serol o g i c a l betes (44–48,108, 112–114). Despite the fact risk (90). The number of patients with dia- responses (e.g., type specific antibody that the administration of the vaccine in these betes in other case series of pneumococcal responses of twofold over baseline, similar to studies was according to the pref e r ences of infection has been as little as 1–3% in patient no r mal control subjects, and a postvaccina- individual providers and patients during the gr oups 50 years of age and 12–19% in tion geometric mean antibody concentration course of clinical care, matching control sub- older groups (16,91–95). Several rev i e w s 300 ng/ml). jects for comorbid conditions (in case-control

DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 103 Technical Review series) and logistical analysis for potentially despite the uncertainty of pneumococcal vac- sions or discharge plans (88,115). Addi- confounding variables such as race, eco- ci n e ’ s efficacy in nonbacteremic illness. tional health system processes that have nomic status, prior hospitalization, and prior helped facilitate improved vaccination rates receipt of immunization suggests that a sys- VACCINE IMPLEMENTATION have included empowered nursing staff with tematic bias would have been an unlikely S T R AT E G I E S — The ACIP re c o m- standing orders to vaccinate, standardi z e d explanation to explain the authors’ observa - mends that individuals with diabetes rec e i v e i n f o rmation and documentation form s tions (48,112). Shapiro et al. (48) rep o r ted a at least one lifetime immunization with (paper and electronic), targeted educational pro s p ective study over 6 years in an attempt pneumococcal vaccine and annual influenza eff o r ts to patients and providers, special to control for these biases and demonstrated vaccinations (88,115). An immunization immunization programs throughout the vaccine efficacy in their case-control study rate of at least 60% is a national health objec- ye a r , notification, and mailings (126–133). and indirect cohort analysis. Additional tive (116). The Health Care Financing Although case management has been used rep o r ts demonstrate clinical efficacy in at- Administration has encouraged collabora- to improve immunization rates, the cost is risk population groups, including grou p s tion on immunization quality improv e m e n t significant (134). Tar geting patients attend- w h e re 4–15% of subjects had diabetes p rojects among health care plans and ing specialty clinics for diabetes care can (1 1 3 , 1 1 4 ) . pr oviders. Financial issues have been raised pr ovide a simple and effective method to Fo r rester et al. (47) failed to demon- as a significant barrier in effective immu- pr ovide routine immunization and moni- strate efficacy of the pneumococcal vaccine nization (117). However, when countries toring without major systems ree n g i n e e r - using both an indirect cohort and case-con- have provided free immunization to targe t e d ing. Although there are no rep o r ts reg a rd i n g tr ol design. Bias in this study included lim- gr oups, this has not always translated into the value of influenza vaccination for the iting the documentation of vaccination success (58). Despite coverage for the health care team, specifically in caring for status to what was rec o r ded only in the hos- Me d i c a r e population, 65–80% of a hospi- people with diabetes, this should also be pital rec o r d, missing medical rec o r ds in 7% talized cohort (1994–1995) missed oppor- c o n s i d e red as part of a health systems of the original cohort, 30% of cases having tunities for influenza and pneumococcal ap p r oach to preventive services and is rec - a p p a rent nosocomial infections, and a vaccination (118). In this group of patients, ommended by the ACIP. higher incidence of bacteremia in cases com- 67% (95% CI 64.7–68.4) were identified as Knowledge and attitudes of patients and pa r ed with control subjects (18 vs. 7%). In having at least one chronic condition (to pr oviders as well as health system proc e s s e s addition, the analysis was limited to 89 include diabetes). and barriers influence vaccination rates patients of which 23% would be considered In the last 10 years, the concepts and (126,135,136). Individuals who believe that immune incompetent patients (e.g., having tools of continuous quality improv e m e n t influenza and pneumococcal disease are asplenia, dysglobulinemia, renal transplan- (CQI) have been used to improve compli- serious illnesses and vaccination is effe c t i v e tation, nephrotic syndrome, hematological ance rates for preventive services, including and safe are more likely to be immunized if malignancies, and drug-induced immuno- immunization (119,120). Despite this effo rt , they are advised to do so by their prov i d e r su p p r ession). The “medium-risk patient immunization rates remain unacceptably (6,126). Provider recommendation is a gr oup” (which included patients with dia- lo w , whereas systematic approaches to pre- highly significant and independent pred i c t o r betes) accounted for only 64% of the total ventive services have become more and more of successful immunization (126). co h o r t of patients. Independent of concerns complex. Improving immunization rates for reg a r ding the power to detect vaccine effi - people with diabetes is only part of the large r Conclusion cacy in these patients, there were no specific health system’s difficulties in implementing Pr ovider recommendation is a cost-effe c t i v e details concerning the patients with dia- pr eventive services and guidelines for man- immunization implementation strategy. betes. Because of these issues, it is not pos- agement of chronic diseases. The number of Health system processes and CQI can sup- sible to draw conclusions about patients overlapping CQI effo r ts can be overwh e l m - po r t the provider and patient in this and with diabetes. ing, even for the most motivated prov i d e r s other effective implementation strategies. It has been suggested that pneumococ- and patients. A user-friendly tracking system Although data collection and tracking cal vaccine might be effective in prev e n t i n g is essential to achieve successful prioritiza- appears essential in effective implementation ba c t e r emic pneumococcal disease but not tion, scheduling, and rec o r d keeping. strategies, targeting at-risk groups in sub- other forms of pneumococcal infection (26). A number of paper and paper/computer specialty clinics and during hospitalizations Although this could be unique to the vac- databases have assisted in implementation can greatly simplify this process and trans- cine, it might also be because of the diffi c u l t y strategies (10,121,122). Timely rep o r ting to late into significant cost savings and the pre- in making a definitive clinical diagnosis of pr ovider and patient is essential for success vention of disease. pneumococcal infection without documen- in using these implementation strategies tation of bacteremia. Either explanation (121). Previous hospitalization has been CO N C L U S I O N S — Individuals with might account for lower efficacy rates in identified as a risk factor for subsequent diabetes have at least the same, if not studies rep o r ting nonbacteremic pres e n t a - serious influenza and pneumococcal infec- in c r eased, risk of influenza and pneumo- tion of pneumococcal infection. tion (123–125). Based on the observa t i o n coccal infection as other patients with Conclusion. Many studies have shown that that hospitalization may be a marker for this ch r onic diseases. In addition, patients with the vaccine is effective in reducing pneumo- in c r eased risk, the ACIP and the American diabetes often have associated comorbid coccal bacteremia, a common complication Hospital Association have encouraged conditions that increase this risk. Certa i n of pneumococcal infection with a high mor- obtaining vaccination histories from all inpa- aspects of their care and the acute and tality rate (46,47,105,106,112). This effi c a c y tients and have suggested that vaccinations c h ronic complications of their disease alone supports its use in people with di a b e t e s , be implemented as part of prolonged admis- appear to uniquely increase their risk for

104 DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 Smith and Poland these infections. While specific acquired and monia. Ar ch Intern Med 59:602–619, 1937 be t o l o g i a 30:132–143, 1987 genetic defects in immune surveillance may 14 . Kramer MR, Rudensky B, Hadas-Halperin I, 29 . He r old KC, Huen H, Gould L, Traisman H, in c r ease the risks for infection and the lack Isacsohn M, Melzer E: Pneumococcal bac- Rubenstein AH: Alterations in lymphocyte of robust immune response to immuniza- te r emia: no change in mortality in 30 years: sub populations in type 1 (insulin-depen- tion, it appears that immunization in this at- analysis of 104 cases and review of the lit- dent) diabetes mellitus: exploration of pos- er a t u r e. Isr J Med Sci 23:174–180, 1987 risk patient population is effective. Effe c t i v e sible mechanisms and relationships to 15 . Austrian R, Gold J: Pneumococcal bac- autoimmune phenomena. Di a b e t o l o g i a 27 : implementation strategies for immunization t e remia with special re f e rence to bac- 102–105, 1984 in this patient population appear justified. te r emic pneumococcal pneumonia. An n 30 . Katz S, Klein B, Elian I, Fishman P, In t e r n Med 60:759–776, 1964 Djaldetti M: Phagocytotic activity of mono- 16 . Perlino CA, Rimland D: Alcoholism, leukope- cytes from diabetic patients. Diabetes Care ni a , and pneumococcal sepsis. Am Rev Respir Re f e re n c e s 6:479–482, 1983 Di s 132:757–760, 1985 1. Bu r row GN, Hazlett BE, Philips MJ: A case 31 . Nolan CM, Beaty HN, Bagdade JD: Furth e r 17 . Landesman SH, Schiffman G: Assessment of diabetes mellitus. N Engl J Med 30 6 : characterization of the impaired bacterici- of the antibody response to pneumococcal 340–343, 1982 dal function of granulocytes in patients vaccine in high risk populations. Rev Infect 2. Diepersloot RJA, Bouter KP, Hoekstra JBL: with poorly controlled diabetes. Di a b e t e s Di s 3 (Suppl.):S184–S196, 1981 Influenza infection and diabetes mellitus: 27:889–894, 1978 18 . Musher DRM, Chapman AJ, Goree A, Jon- case for annual vaccination. Diabetes Care 32 . Bybee JD, Rogers DE: The phagocytic sson S, Briles D, Baughn RE: Natural and 13:876–882, 1990 activity of polymorphonuclear leukocytes va c c i n e - r elated immunity to Strep t o c o c c u s 3. Ha n d w e r ger BS, Fernandes G, Brown DM: obtained from patients with diabetes mel- pneumoniae. J Infect Dis 15 4 : 2 4 5 – 2 5 6 , Immune and autoimmune aspects of dia- litus. J Lab Clin Med 64:1–13, 1964 19 8 6 betes mellitus. Hum Pathol 11 : 3 3 8 – 3 5 2 , 33 . Rayfield EJ, Ault MJ, Keusch GT, Brot h e r s 19 . Sc h w a r tz JS: Pneumococcal vaccine: clinical 19 8 0 MJ, Nechemias C, Smith H: Infection and 4. Nicholson KG, Snacken R, Palache AM: ef ficacy and effectiveness. Ann Intern Med diabetes: the case for glucose control. Am J Influenza immunization policies in Europ e 96:208–220, 1982 Me d 72:439–450, 1982 and the United States. Vac c i n e 13 : 3 6 5 – 3 6 7 , 20 . Giebink GS, Le CT, Cosio FG, Spika JS, 34 . Chandler PT, Chandler SD: Pathogenic 19 9 5 Sc h i f fman G: Serum antibody responses in ca r rier rate in diabetes mellitus. Am J Med 5. The National Vaccine Advisory Committee high risk children and adults to vaccina- Sc i 273:259–265, 1977 of the National Vaccine Program: Adult tion with capsular polysaccharides of 35 . Finn PD, Holden FA: Observations and immunization: a rep o r t by the National St r eptococcus pneumoniae. Rev Infect Dis comments concerning the isolation of grou p Vaccine Advisory Committee. Atlanta, GA, 3:S168–S178, 1981 B b-hemolytic streptococci from human U.S. Department of Health and Human 21 . Beam TR, Crigler ED, Goldman JK, Schiff- so u r ces. Can Med Assoc J 10 3 : 2 4 9 – 2 5 2 , Se r vices, 1994 man GL: Antibody response to polyvalent 19 7 0 6. N g u y e n - Va n - Tam JS, Nicholson KG: pneumococcal polysaccharide vaccine in 36 . Casey JI, Maturio S, Albin J, Edberg SC: Influenza immunization: vaccine off e r, diabetics. JA M A 244:2621–2624, 1980 Comparison of carriage rates of group B request and uptake in high-risk patients 22 . L e d e rman MM, Schiffman G, Rodman St r eptococcus in diabetic and nondiabetic during the 1991/2 season. Epidemiol Infect HM: Pneumococcal immunization in adult persons. Am J Epidemiol 1 1 6 : 7 0 4 – 7 0 8 , 111:347–355, 1993 diabetics. Di a b e t e s 30:119–121, 1981 19 8 2 7. Ho M, Marger M, Beart J, Yip I, Shekelle P: 23 . Kaneshige H: Nonenzymatic glycosylation 37 . Watkins PJ, Soler NG, Fitzgerald MG, Is the quality of diabetes care better in a dia- of serum IgG and its effect on antibody Malins JM: Diabetic ketoacidosis during betes clinic or in a general medicine clinic? activity in patients with diabetes mellitus. the influenza epidemic. Br Med J 4: 8 9 – 9 1 , Diabetes Care 20:472–475, 1997 Di a b e t e s 36:822–828, 1987 19 7 0 8. Peters AL, Legorreta AP, Ossorio RC, 24 . Diepersloot RJA, Bouter KP, Beyer WEP, 38 . Bouter KP, Diepersloot RJA, van Romunde Davidson MB: Quality of outpatient care Hodkstra JBL, Masurel N: Humoral LKJ, Uitslager R, Masurel N, Hoekstra JBL, pr ovided to diabetic patients. Diabetes Care immune response and delayed type hyper- Erkelens DW: Effect of epidemic influenza 19:601–605, 1996 sensitivity to influenza vaccine in patients on ketoacidosis, pneumonia and death in 9. Marshall CL, Bluestein M, Chapin C, Davis with diabetes mellitus. Di a b e t o l o g i a 30 : diabetes mellitus: a hospital register surve y T, Gersten J, Harris C, Hodgin A, Larsen W, 397–401, 1987 of 1976–1979 in the Netherlands. Di a - Ri g b e r g H, Krishnaswami V, Darling B: 25 . Diepersloot RJA, Bouter KP, van Beek R, betes Res Clin Pract 12:61–68, 1991 Outpatient management of diabetes melli- Lucas CJ, Masurel N, Erkelens DW: Cyto- 39 . Glezen WP, Decker M, Perrotta DM: Surve y tus in five Arizona Medicare managed care toxic T-cell response to influenza A subunit of underlying conditions of persons hospi- plans. Am J Med Qual 11:87–93, 1996 vaccine in patients with type 1 diabetes talized with acute res p i r a t o r y disease dur- 10 . Hershey CO, Karuza J: Assessment of pre- mellitus. Neth J Med 35:68–75, 1989 i n g influenza epidemics in Houston, ventive health care: design considerations. 26 . Fe e r y BJ, Hartman LJ, Hampson AW, Proi - 1 9 78 –1981 . Am Rev Resp Dis Pr ev Med 26:59–67, 1997 etto J: Influenza immunization in adults 136:550–555, 1987 11 . Douvin C, Simon D, Charles M-A, Deforge s with diabetes mellitus. Diabetes Care 6: 40 . Sc h w a r zmann SW, Adler JL, Sullivan RJ, L, Bierling P, Lehner V, Budkowska A, 475–478, 1983 Marine WM: Bacterial pneumonia during Dhumeaus D: Hepatitis B vaccination in 27 . Kaye WA, Adri MNS, Soeldner JS, Rabi- the Hong Kong influenza epidemic of diabetic patients. Diabetes Care 20 : 1 4 8 – 1 5 1 , nowe SL, Kaldany A, Kahn CR, Bistrian B, 19 6 8–1969. Ar ch Intern Med 12 7 : 1 0 3 7 – 19 9 7 Srikanta S, Ganda OP, Eisenbarth GS: 1041, 1971 12 . Casey JI: Host defense abnormalities in dia- Ac q u i r ed defect in interleukin-2 prod u c - 41 . Bisno AL, Griffin JP, Van Epps KA, Niell betic patients. In Diabetes Mellitus. Vol 5. tion in patients with type 1 diabetes melli- HB, Rytel MW: Pneumonia and Hong Rifkin H, Raskin P, Eds. Bowie, MD, Robert tus. N Engl J Med 315:920–924, 1986 Kong influenza: a prospective study of the J. Brady Company, 1981, p. 219–223 28 . Dr ell DW, Notkins AL: Multiple immuno- 1968–1969 epidemic. Am J Med Sci 26 1 : 13 . Tilghman RC, Finland M: Clinical signifi- logical abnormalities in patients with type 1 251–263, 1971 cance of bacteremia in pneumococcal pneu- (insulin dependent) diabetes mellitus. Di a - 42 . Ficicioglu C, Mikla S, Midilli K, Aydin A,

DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 105 Technical Review

Cam H, Ergin S: Reduced immune res p o n s e tality from epidemic influenza, 1957–1966. sponses to killed influenza vaccine in to hepatitis B vaccine in children with Am J Epidemiol 100:40–48, 1974 patients with type 1 diabetes: altere d insulin dependent diabetes. Acta Paediatr 56 . Ca m e r on AS, Roder DM, Esterman AJ, response associated with HLA-D3 and DR4. Jp n 37:687–690, 1995 Mo o r e BW: Mortality from influenza and J Lab Clin Med 112:595–602, 1988 43 . Wismans PJ, van Hattum J, de Gast GC, allied infections in South Australia during 70 . Hiltunen M, Hyoty H, Leinikki P, Bouter KP, Diepersloot RJA, Maikoe TJ, 1969–1981. Med J Aust 142:14–17, 1985 Akerblom HK, Tuomilehto J, Vesikari T: Mudde GC: A prospective study of in vitro 57 . McBean AM, Babish JD, War ren JL, Melson Low mumps antibody levels induced by anit-HBs producing B cells (Spot-ELISA) EA: The effect of influenza epidemics on m u m p s - m e a s l e s - rubella vaccinations in following primary and supplementary vac- the hospitalization of persons 65 years and type 1 diabetic children. Diabet Med 11 : cination with a recombinant hepatitis B ol d e r . In Options for the Control of Influenza 942–946, 1994 vaccine in insulin dependent diabetic II . Hannoun C, Klenk HD, Kendal AP, 71 . El-Reshaid K, Al-Mufti S, Johny KV, patients and matched controls. J Med Viro l Ruben FL, Eds. Amsterdam, Excerpta Sugathan TN: Comparison of two immu- 35:216–222, 1991 Medica, 1993, p. 25–37 nization schedules with re c o m b i n a n t 44 . B roome CV, Facklam RR, Fraser DW: 5 8 .C a rrat F, Val l e r on A-J: Influenza morta l i t y hepatitis B vaccine and natural immunity Pneumococcal disease after pneumococcal among the elderly in France, 1980–90: ac q u i r ed by hepatitis B infection in dialysis vaccination: an alternative method to esti- how many deaths may have been avoided patients. Vac c i n e 12:223–228, 1994 mate the efficacy of pneumococcal vac- th r ough vaccination? J Epidemiol Comm 72 . Bouter KP, Diepersloot RJA, Wismans PJ, cine. N Engl J Med 303:549–552, 1980 He a l t h 49:419–425, 1995 Gmelig Meyling FHJ, Hoekstra JBL, Hei- 45 . Butler JC, Breiman RF, Campbell JF, Lip- 59 . Barker WH, Mullooly JP: Pneumonia and jtink RA, van Hattum J: Humoral immune man HB, Broome CV, Facklam FR: Pneu- influenza deaths during epidemics: impli- response to a yeast-derived hepatitis B vac- mococcal polysaccharide and vaccine cations for prevention. Ar ch Intern Med cine in patients with type 1 diabetes melli- ef ficacy: an evaluation of current rec o m - 142:85–89, 1982 tus. Diabet Med 9:66–69, 1992 mendations. JA M A 270:1826–1831, 1993 60 . Alling DW, Blackwelder WC, Stuart- H a rr i s 73 . Fedson DS, Wajda A, Nicol JP, Hammond 46 . Bolan G, Broome CV, Facklam RR, CH: A study of excess mortality during GW , Kaiser DL, Roos LL: Clinical effe c t i v e - Plikaytis BD, Fraser DW, Schlech WE III: influenza epidemics in the United States, ness of influenza vaccination in Manitoba. Pneumococcal vaccine efficacy in selected 1968–1976. Am J Epidemiol 11 3 : 3 0 – 4 3 , JAMA 270:1956–1961, 1993 populations in the United States. A n n 19 8 1 74 . Foster DA, Talsma A, Furumoto-Dawson A, In t e r n Med 104:1–6, 1986 61 . N g u y e n - Va n - Tam JS, Nicholson KG: Ohmit SE, Margulies JR, Arden NH, Monto 47 . Fo r rester HL, Jahnigen DW, LaForce FM: Influenza deaths in Leicestershire during AS: Influenza vaccine effectiveness in pre- In e f ficacy of pneumococcal vaccine in a the 1989–90 epidemic: implications for venting hospitalization for pneumonia in high-risk population. Am J Med 83 : 4 2 5 – pr evention. Epidemiol Infect 10 8 : 5 3 7 – 5 4 5 , the elderly. Am J Epidemiol 13 6 : 2 9 6 – 3 0 7 , 430, 1987 19 9 2 19 9 2 48 . Sh a p i r o ED, Berg AT, Austrian R, Schroe d e r 62 . Goslings WRO, Mulder J, Djajadiningrat J, 75 . Gr oss PA, Quinnan GV, Rodstein M, LaM- D, Parcells V, Margolis A, Adair RK, Ma s u r el N: Staphylococcal pneumonia in ontagne JR, Kaslow RA, Saah AJ, Wal l e n - Clemens JD: The protective efficacy of poly- influenza: in relation to antecedent staphy- stein S, Neufeld R, Denning C, Gaerlan P: valent pneumococcal polysaccharide vac- lococcal skin infection. La n c e t ii : 4 2 8 – 4 3 0 , Association of influenza immunization ci n e . N Engl J Med 325:1453–1460, 1991 19 5 9 with reduction in mortality in an elderly 49 . Nichol KL, Margolis KL, Wuo r enma J, Von 63 . Winterbauer RH, Ludwig WR, Hammar population: a prospective study. Ar ch Intern St e rn b e r g T: The efficacy and cost effe c - SP: Clinical course, management , and long Me d 148:562–565, 1988 tiveness of vaccination against influenza te r m sequelae of res p i r a t o r y failure due to 76 . Edmondson WP, Rothenberg R, White PW, among elderly persons living in the com- influenza viral pneumonia. Johns Hopkins Gwaltney JM: A comparison of subcuta- mu n i t y . N Engl J Med 331:778–784, 1994 Med J 141:148–155, 1977 neous, nasal, and combined influenza vac- 50 . Strikas R, Cook P, Kuller L, Talbott E, 64 . Giles C, Shuttleworth EM: Post-morte m cination. II. Protection against natural P l o u ffe J, File T, Poggio E, Kidder D, findings in 46 influenza deaths. La n c e t ii : challenge. Am J Epidemiol 93 : 4 8 0 – 4 8 6 , Baughman A, Manoharan A, Hutton E: 1224–1226, 1957 19 7 1 Case control study in Ohio and Pennsyl- 65 . Sabin AB: Mortality from pneumonia and 77 . St u a r t WH, Dull B, Newton LH, McQueen vania on prevention of hospitalization by risk conditions during influenza epi- JL, Schiff ER: Evaluation of monovalent influenza vaccination. In Options for the de m i c s . JAMA 237:2823–2828, 1977 influenza vaccine in a ret i r ement commu- Co n t r ol of Influenza II. Hannoun C, Klenk 66 . McElhaney JE, Pinkoski MJ, Au D, Lechelt nity during the epidemic of 1965–66. HD, Kendal AP, Ruben FL, Eds. Amster- KE, Bleackley RC, Meneilly GS: Helper JA M A 209:232–238, 1969 dam, Excerpta Medica, 1993, p. 153–160 and cytotoxic T lymphocyte responses to 78 . Pa t r i a r ca PA, Weber JA, Parker RA, Hall 51 . Ma r tin WJ: Recent changes in the death influenza vaccination in healthy compared WN, Kendal AP, Bregman DJ, Schonberge r rate from influenza. Br Med J 1: 2 6 7 – 2 6 8 , to diabetic elderly. Vac c i n e 14 : 5 3 9 – 5 4 4 , LB: Efficacy of influenza vaccine in nursing 19 5 0 19 9 6 homes: reduction in illness and complica- 52 . Stocks P, Camb MD: Influenza epidemics 67 . Possilli P, Gale WAM, Visalli N, Baroni M, tions during an influenza A (H3N2) epi- on the certified causes of death. La n c e t ii : Cr ovari P, Frighi V, Cavallo MG, Andrea n i demic. JA M A 252:1136–1139, 1985 386–395, 1935 D: The immune response to influenza vac- 79 . St r a s s b u r g MA, Greenland S, Sorvillo FJ, 53 . E i c k h o ff TC, Sherman IL, Serfling RE: cination in diabetic patients. Di a b e t o l o g i a Lieb LE, Habel LA: Influenza in the elderly: Ob s e r vations on excess mortality associ- 29:850–854, 1986 rep o r t of an outbreak and a review of vac- ated with epidemic influenza. JA M A 17 6 : 68 . Scheinin T, Tran-Minh N-N, Kontiainen S: cine effectiveness rep o r ts. Vac c i n e 4: 3 8 – 4 4 , 104–110, 1961 T cell responses to PPD in BCG- vaccinated 19 8 6 54 . P e t e r s d o rf RG, Fusco JJ, Harter DH, ch i l d r en with insulin-dependent diabetes 80 . Gr oss PA, Hermogenes AW, Sacks HS, Lau Albrink WS: Pulmonary infections com- mellitus and controls. Acta Paediatr 83 : J, Levandowski RA: The efficacy of plicating Asian influenza. Ar ch Intern Med 337–338, 1994 influenza vaccine in elderly persons: a 103:262–272, 1959 69 . Ruben FL, Fireman P, LaPorte RE, Drash meta-analysis and review of the literature. 55 . Ho u s w o r th J, Langmuir AD: Excess mor- AL, Uhrin M, Ve rgona R: Immune re - Ann Intern Med 123:518–527, 1995

106 DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 Smith and Poland

81 . Barker WH, Mullooly JP: Influenza vacci- c o m m u n i t y - a c q u i red and nosocomial erative study. N Engl J Med 31 5 : 1 3 1 8 – nation of elderly persons: reduction in pneumococcal bacteremia. Am Rev Respir 1327, 1986 pneumonia and influenza hospitalizations Di s 123:265–268, 1981 10 9 . Sh a p i r o ED: Pneumococcal vaccine fail- and deaths. JA M A 244:2547–2549, 1980 94 . Wat a n a k u n a k o r n C, Greifenstein KS, Jar- ur e. N Engl J Med 316:1272–1273, 1987 82 . Saah AJ, Neufeld R, Rodstein M, Montague joura DG, Blend D, Cugino A, Ognibene 11 0 . Hirschman JV, Lipsky BA: Pneumococcal JR, Blackwelder WC, Gross P, Quinnan G, AJ: Pneumococcal bacteremia in thre e vaccine in the United States: a critical Kaslow RA: Influenza vaccine and pneu- community teaching hospitals from 1980 analysis. JA M A 246:1428–1432, 1981 monia mortality in a nursing home popula- to 1989. Ch e s t 103:1152–1156, 1993 11 1 . Fraser DR, Broom CV: Pneumococcal vac- tion. Ar ch Intern Med 146:2353–2357, 1986 95 . A l v a rez S, Guarderas J, Shell CG, cine: to use or not (editorial). JA M A 24 5 : 83 . Mullooly JP, Bennett MD, Hornb r ook MC, Hotsclaw-Berk S, Berk SL: Nosocomial 498–499, 1981 Barker WH, Williams WW, Patriarca PA, pneumococcal bacteremia. Ar ch Intern Med 11 2 . Sh a p i r o ED, Clemens JD: A control l e d Rhodes PH: Influenza vaccination pro- 146:1509–1512, 1986 evaluation of the protective efficacy of grams for elderly persons: cost-effe c t i v e - 96 . Austrian R: Random gleanings from a life pneumococcal vaccine for patients at high ness in a health maintenance orga n i z a t i o n . with the pneumococcus. J Infect Dis 13 1 : risk of serious pneumococcal infections. Ann Intern Med 121:947–952, 1994 474–484, 1975 Ann Intern Med 104:325–330, 1984 84 . Ng u y e n - Va n - T am JS, Ahmed AH, Nichol- 97 . Br uyn GAW, van der Meer JWM, Herma n s 11 3 . Sims RV, Steinmann WC, McConville JH, son KG, Pearson JCG: Reduction in hospi- J, Knoppert W: Pneumococcal bacterem i a King LR, Zwick WC, Schwartz JS: The clin- tal admissions for pneumonia, influenza, in adults over a 10-year period at Univer- ical effectiveness of pneumococcal vaccine br onchitis and emphysema associated with sity Hospital, Leiden. Rev Infect Dis in the elderly. Ann Intern Med 10 8 : 6 5 3 – 6 5 7 , influenza vaccine during the 1989–90 epi- 10:446–450, 1988 19 8 8 demic in Leicestershire, UK. In Options for 98 . Chang JI, Mylotte JM: Pneumococcal bac- 11 4 . Koivula I, Sten M, Leinonen M, Makela the Control of Influenza II. Hannoun C, Klenk te r emia: update from an adult hospital PH: Clinical efficacy of pneumococcal vac- HD, Kendal AP, Ruben FL, Eds. Amster- with a high rate of nosocomial cases. J Am cine in the elderly: a randomized, single- dam, Excerpta Medica, 1993, p. 107–112 Geriatr Soc 35:747–754, 1987 blind population-based trial. Am J Med 85 . Monto AS, Ohmit SE, Foster DA, Furu- 99 . Ruben FL, Norden CW, Korica Y: Pneu- 103:281–290, 1997 moto-Dawson A, Arden NA: Case-control mococcal bacteremia at a medical/surgi c a l 11 5 . CDC: Prevention and control of influenza: study in Michigan on prevention of hospi- hospital for adults between 1975–1980. recommendations of the Advisory Com- talization by vaccination, 1989–1991. In Am J Med 77:1091–1094, 1984 mittee on Immunization Practices (ACIP). Options for the Control of Influenza II. Han- 10 0 . Filice GA, Darby CP, Fraser DW: Pneumo- MMWR 46:1–25, 1997 noun C, Klenk HD, Kendal AP, Ruben FL, coccal bacteremia in Charleston County, 11 6 . Public Health Service: Healthy People 2000: Eds. Amsterdam, Excerpta Medica, 1993, South Carolina. Am J Epidemiol 11 2 : 8 2 8 – National Health Promotion and Disease Pre - p. 135–141 835, 1980 vention Objectives. Washington, DC, U.S. 86 . Paul WS, Cowan J, Jackson GG: Acute res - 10 1 . Mufson MA, Kruss DM, Wasil RE, Metzger De p a r tment of Health and Human Ser- pi r a t o r y illness among immunized and WI: Capsular types and outcome of bac- vices, Public Health Service, 1991, p. nonimmunized patients with high-risk fac- te r emic pneumococcal disease in the antibi- 122–123 (DHHS publ. no. PHS 91-50213) tors during a split season of influenza A otic era. Ar ch Intern Med 13 4 : 5 0 5 – 5 1 0 , 11 7 . Be r gner L, Yer gy AS: Low income and bar- and B. J Infect Dis 157:633–639, 1988 19 7 4 riers to use of health services. N Engl J Med 87 . Nicholson KG, Stone AJ, Botha JL, Ray- 10 2 . Hook EW, Horton CA, Schaberg DR: Fail- 278:541–546, 1968 mond NT: Effectiveness of influenza vac- ur e of intensive care unit support to influ- 11 8 . CDC: Missed opportunities for pneumo- cine in reducing hospital admissions in ence mortality from pneumococcal coccal and influenza vaccination of people with diabetes. In Options for the ba c t e r emia. JA M A 249:1055–1057, 1983 Me d i c a r e pneumonia inpatients: 12 western Co n t r ol of Influenza III. Brown LE, Hamp- 10 3 . Gransden WR, Eykyn SJ, Phillips I: Pneu- states, 1995. MM W R 46:919–923, 1997 son AW, Webster RG, Eds. Elsevier Sci- mococcal bacteremia: 325 episodes diag- 11 9 . Carlin E, Carlson R, Nordin J: Using con- ence, 1996, p. 113–118 nosed at St Thomas’s Hospital. Br Med J tinuous quality improvement tools to 88 . CDC: Prevention of pneumococcal dis- 290:505–508, 1985 im p r ove pediatric immunization rates. Jt ease: recommendations of the Advisory 1 0 4 .G ruer LD, McKendrick MW, Geddes AM: Comm J Qual Improv 22:277–288, 1996 Committee on Immunization Practices Pneumococcal bacteremia: a continuing 12 0 . Zi m m e r man RK, Ruben FL, Ahwesh MA: (ACIP). MM W R 46:1–24, 1997 challenge. Q J Med 210:259–270, 1984 Influenza, influenza vaccine, and amanta- 89 . S t u a rt - H a rris CH, Franks Z, Ty rrell D: 10 5 . Smit P, Oberholzer D, Hayden-Smith S, dine/rimantadine. J Fam Pract 45 : 1 0 7 – 1 2 4 , Deaths from influenza: a statistical and lab- Ko o r nhof HJ, Hilleman MR: Protective effi - 19 9 7 or a t o r y investigation. Br Med J 1: 2 6 3 – 2 6 6 , cacy of pneumococcal polysaccharide vac- 12 1 . Ba r ton MB, Schoenbaum SC: Improv i n g 19 5 0 cines. JA M A 238:2613–2616, 1977 influenza vaccination perfo r mance in an 90 . Moss JM: Pneumococcus infection in dia- 10 6 . Austrian R, Douglas RM, Shiffman G, HMO setting: the use of computer-g e n e r - betes mellitus: is this a justification for Coetzee AM, Koornhof HJ, Hayden- Smith ated reminders and peer comparison feed- immunization? J A M A 2 4 3 : 2 3 0 1 – 2 3 0 3 , S, Reid RD: Prevention of pneumococcal back. Am J Public Health 80:534–536, 1990 19 8 0 pneumonia by vaccination. Trans Assoc Am 12 2 . Klachko DM, Wright DL, Gardner DW: 91 . Fedson DS, Chiarello LA: Previous hospi- Ph y s i c i a n s 89:184–189, 1976 Ef fect of a microc o m p u t e r -based reg i s t r y tal care and pneumococcal bacterem i a : 10 7 . Sc h w a r tz JS: Pneumococcal vaccine: clini- on adult immunizations. J Fam Pract 29 : im p o r tance for pneumococcal immuniza- cal efficacy and effe c t i v e n e s s . Ann Intern 169–172, 1989 tion. Ar ch Intern Med 143:885–889, 1983 Me d 96:208–220, 1982 12 3 . Fedson DS: Improving the use of pneu- 92 . Finkelstein MS, Petkun WM, Free d m a n 1 0 8 .S i m b e r k o ff MS, Cross AP, Al-Ibrahim M, mococcal vaccine through a strategy of ML, Antopol SC: Pneumococcal bac- Baltch AL, Geiseler PJ, Nadler J, Rich- hospital-based immunization: a review of te r emia in adults: age-dependent diffe r - mond AS, Smith RP, Schiffman G, Shepard its rationale and implications. J Am Geri - ences in presentation and in outcome. J Am DS, Van Eeckhout JP: Efficacy of pneu- atric Soc 33:142–150, 1985 Geriatr Soc 31:19–27, 1983 mococcal vaccine in high-risk patients: 12 4 . Fedson DS, Wajda A, Nicol JP, Roos LL: 93 . Mylotte JM, Beam TR Jr: Comparison of results of a veterans administration coop- Disparity between influenza vaccination

DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000 107 Technical Review

rates and risks for influenza-associated hos- 13 0 . McDowell I, Newell C, Rosser W: Com- 13 6 . Fiebach N, Beckett W: Prevention of res - pital discharge and death in Manitoba in parison of three methods of re c a l l i n g pi r a t o r y infections in adults: influenza and 1982–1983. Ann Intern Med 11 6 : 5 5 0 – 5 5 5 , patients for influenza vaccination. Can Med pneumococcal vaccines. Ar ch Intern Med 19 9 2 Assoc J 135:991–997, 1986 154:2545–2557, 1994 12 5 . Barker WH, Mullooly JP: Pneumonia and 13 1 . Mullooly JP: Increasing influenza vaccina- 13 7 . Finland M, Peterson OL, Strauss E: influenza deaths during epidemics: impli- tion among high-risk elderly: a randomized Staphylococcal pneumonia occurring dur- cations for prevention. Ar ch Intern Med co n t r olled trial of a mail cue in an HMO ing an epidemic of influenza. Ar ch Intern 142:85–89, 1992 setting. Am J Public Health 77 : 6 2 6 – 6 2 7 , Me d 70:183–205, 1942 12 6 . Nichol KL, MacDonald RM, Hauge M: Fac- 19 8 7 13 8 . Ma r tin CM, Kunin CM, Gottlieb LS, Barne s tors associated with influenza and pneu- 13 2 . Cr ouse BJ, Nichol K, Peterson DC, Grimm MW , Liu C, Finland M: Asian influenza A in mococcal vaccination behavior among MB: Hospital-based strategies for improv - Boston, 1957–1958: observations in thirty - high-risk adults. J Gen Intern Med 1 1 : ing influenza vaccination rates. J Fam Pract two influenza-associated fatal cases. Arc h 673–677, 1996 38:258–261, 1994 In t e r n Med 103:515–531, 1959 12 7 . Cheney C, Ramsdell JW: Effect of medical 13 3 . Fedson DS, Kessler HA: A hospital-based 1 3 9 .M a rtin CM, Kunin CM, Gottlieb LS, Fin- rec o r ds’ checklists on implementation of influenza immunization program, 1977– land M: Asian influenza A in Boston, periodic health measures. Am J Med 78. Am J Public Health 73:442–445, 1983 1957–1958: severe staphylococcal pneu- 83:129–136, 1987 13 4 . Wood D, Halfon N, Donald-Sherbourne C, monia complicating influenza. Ar ch Intern 12 8 . Ti e rney WM, Hui SL, McDonald CJ: Mazel RM, Schuster M, Hamlin JS, Perey r a Me d 103:532–542, 1959 Delayed feedback of physician perfo rm a n c e M, Camp P, Grabowsky M, Duan N: 14 0 . Oseasohn R, Adelson L, Kaji M: Clinico- versus immediate reminders to perfo r m pre- I n c reasing immunization rates among pathologic study of thirty - t h r ee fatal cases of ventive care. Med Care 24:659–666, 1986 in n e r- c i t y , African American children: a Asian influenza. N Engl J Med 26 0 : 5 0 9 – 5 1 8 , 12 9 . Win i c k o f f RN, Coltin KL, Morgan MM, randomized trial of case management. 19 5 9 Buxbaum RC, Barnett GO: Impro v i n g JA M A 279:29–34, 1998 14 1 . Neuzil KM, Reed GW, Mitchel EF Jr, Grif- physician perfo r mance through peer com- 13 5 . Fedson DS: Influenza and pneumococcal fin MR: Influenza-associated morbidity parison feedback. Med Care 22 : 5 2 7 – 5 3 4 , immunization strategies for physicians. and mortality in young and middle-aged 19 8 4 Ch e s t 91:436–443, 1994 women. JA M A 281:901–907, 1999

108 DIABETES CARE, VOLUME 23, NUMBER 1, JANUARY 2000