Evaluation of new external quality assessment guidelines involving random blinded rechecking of acid-fast bacilli smears in a pilot project setting in Mexico.

SETTING: Laboratories in Mexico that support the national tuberculosis (TB) control program have been involved in an acid-fast bacilli (AFB) microscopy external quality assurance program which includes rechecking 100% of smears identified as AFB-positive by the local laboratories and 10% of smears identified as AFB-negative. Very few errors have been detected in Mexico using non-random selection and unblinded rechecking of the slides. OBJECTIVE: To evaluate the results from a 1-year pilot program involving blinded rechecking of randomly selected AFB slides from local TB laboratories in two Mexican states and determine its feasibility for future implementation. DESIGN: To reduce potential bias, laboratory staff from the National TB Laboratory, Institute for Epidemiological Diagnosis and Reference (InDRE), performed quarterly statistical sampling of AFB smears and on-site evaluations in local laboratories in each state. AFB smears were rechecked at the respective state laboratories with discordant results resolved at InDRE. RESULTS: A significantly greater percentage of errors was detected on the randomly selected, blinded AFB smears than on the non-randomly selected, unblinded smears. CONCLUSION: Random blinded rechecking provides more accurate estimates of AFB microscopy results, resulting in improved diagnosis and monitoring of treatment response.

Implementation of proficiency testing in conjunction with a rechecking system for external quality assurance in tuberculosis laboratories in Mexico.

SETTING: In developing countries, tuberculosis is diagnosed by identification of acid-fast bacilli (AFB) on sputum smears. OBJECTIVE: To evaluate the quality of AFB microscopy, the Mexican Secretary of Health National Reference Laboratory implemented proficiency testing for its network of 637 laboratories. DESIGN: A total of 586 (92%) laboratories were inspected and 430 technicians evaluated by proficiency testing consisting of 10 slides with known numbers of AFB. Results were compared with those of slide rechecking and with proficiency testing performed 2 years later. RESULTS: Of the 430 technicians evaluated by proficiency testing in 1998, 196 (46%) scored less than 80% and received intensive training in 1999. From a previous mean score of 65% their results increased to 90% (P < 0.0001). In 2001, they again underwent proficiency testing, and the mean score was 83%. The main factors affecting proficiency testing results were the type of laboratory in which the microscopists worked and the number of low-positive slides (1-9/100) in the test. Laboratories whose work was rechecked had better scores (P = 0.002). Proficiency testing scores and the estimated sensitivity of the microscopist's laboratory were associated (P = 0.01). CONCLUSION: External quality assessment and training improve diagnostic performance. Rechecking and proficiency testing are both viable measures of laboratory performance.

Improvement in physician's office laboratory practices, 1989-1994.

BACKGROUND: In 1986 and 1989, the Centers for Disease Control and Prevention sponsored institutes on Critical Issues in Health Laboratory Practice. It was noted during the institutes that physician's office laboratories were a rapidly emerging site for clinical laboratory testing, yet no comprehensive data were available regarding the practice of clinical laboratory medicine in physician's office laboratories. As a mechanism to begin addressing this void, the Centers for Disease Control and Prevention added questions on clinical laboratory practice to the National Ambulatory Medical Care Survey, a national probability sample of ambulatory care provided by office-based physicians. Data were collected for survey years 1989, 1991, 1993, and 1994. METHODS: Each survey was conducted among a nationally representative, random sample of office-based physicians who provide ambulatory patient care. Sample physicians were enlisted using both mail and telephone contacts. Clinical laboratory data were obtained via telephone by trained field representatives. Weighted univariate and multivariate analyses were performed on responses from each of the 4 survey years. Analyses were repeated after combining survey responses from years 1989 and 1991 and 1993 and 1994 as representative of physician's office laboratory practices before and after implementation of the Clinical Laboratory Improvement Amendments of 1988 (CLIA '88) final rule in 1992. RESULTS: Quality laboratory practice indicators showed significant increases during the study interval, with implementation of the CLIA '88 final rule in 1992 playing a pivotal role. Relative to 1992, enrollment in proficiency testing programs increased from 32.4% to 52.7% (P<.001), use of daily quality control samples increased from 79.2% to 89.0% (P<.001), and use of daily quality control with written instructions for action following a questionable quality control result (quality control with action step documentation) increased from 62.6% to 77.2% (P<.001). The presence of a medical technologist or technician in the office laboratory was also significantly and independently associated with each of the quality indicators. Although the percentage of physician's offices performing on-site testing decreased from 56% to 45% during the survey interval, overall testing volume appeared unchanged. CONCLUSIONS: The quality of clinical laboratory practice in physician's office laboratories improved during the study interval (1989-1994) as measured by the quality indicators used in the study. The association of this improvement with implementation of the CLIA '88 final rule and the presence of a trained laboratory professional in the testing site indicate the importance of minimum practice standards and professional expertise in ensuring use of quality laboratory practices. Overall test volume appeared to be stable despite a decreased proportion of physician's offices at which on-site testing was performed.

Laboratory performance in HTLV-I/II analysis.

BACKGROUND: Since 1989, the CDC's Model Performance Evaluation Program has shipped samples to voluntary participant laboratories that test for HTLV antibodies. Each laboratory tests the well-characterized samples, reports the results, and provides information about its testing practices. The data from 15 performance survey periods are reported here. STUDY DESIGN AND METHODS: Multiple logistic regression was used to analyze all data from 15 survey periods from 1989 through 1996. RESULTS: The mean analytic sensitivity for EIA was 99.2 percent per survey period (range, 96-100%), the mean analytic specificity was 97.8 percent (75.6-100%), and the overall accuracy was 88.8 percent (63.8-100%). The mean analytic sensitivity for Western blot was 88.8 percent (75.6-100%); the mean analytic specificity was 95.7 percent (86.7-100%), and the overall accuracy was 91.1 percent (78.1-100%). CONCLUSIONS: Statistical analyses suggested associations between performance and both the retroviral serologic status of the sample and the analytical testing method. Western blot accuracy was associated with weekly testing volume. In early survey periods, performance problems were noted in the analysis of samples from donors with concomitant HTLV and HIV infections and those from donors who were positive for HTLV-II. Technological developments in test methods, such as the addition of recombinant antigens, appeared to have improved the laboratory performance of specific testing methods.

Estimating the linear analytic range.

Although many methods have been proposed to verify assay linearity, or to test for nonlinearity, there are few, if any, statistically sound methods to establish an assay's linear range. In this article, we propose a simple and statistically sound method for initially establishing an assay's linear range as a paradigm for standardization of manufacturers' claims of assay linearity. Simulations verify that the method outperforms using (Pearson's) coefficient of determination (r2) to estimate the linear range, consistently yielding estimates for the linear range which are more accurate. In addition, the method permits the addition of tolerance limits when a certain amount of nonlinearity is acceptable clinically.

A community outbreak of Legionnaires' disease linked to hospital cooling towers: an epidemiological method to calculate dose of exposure.

BACKGROUND: From July to September 1994, 29 cases of community-acquired Legionnaires' disease (LD) were reported in Delaware. The authors conducted an investigation to a) identify the source of the outbreak and risk factors for developing Legionella pneumophila serogroup 1 (Lp-1) pneumonia and b) evaluate the risk associated with the components of cumulative exposure to the source (i.e. distance from the source, frequency of exposure, and duration of exposure). METHODS: A case-control study matched 21 patients to three controls per case by known risk factors for acquiring LD. Controls were selected from patients who attended the same clinic as the respective case-patients. Water samples taken at the hospital, from eight nearby cooling towers, and from four of the patient's homes were cultured for Legionella. Isolates were subtyped using monoclonal antibody (Mab) analysis and arbitrarily primed polymerase chain reaction (AP-PCR). RESULTS: Eleven (52%) of 21 case-patients worked at or visited the hospital compared with 17 (27%) of 63 controls (OR 5.0, 95% CI : 1.1-29). For those who lived, worked, or visited within 4 square miles of the hospital, the risk of illness decreased by 20% for each 0.10 mile from the hospital; it increased by 80% for each visit to the hospital; and it increased by 8% for each hour spent within 0.125 miles of the hospital. Lp-1 was isolated from three patients and both hospital cooling towers. Based on laboratory results no other samples contained Lp-1. The clinical and main-tower isolates all demonstrated Mab pattern 1,2,5,6. AP-PCR matched the main-tower samples with those from two case-patients. CONCLUSION: The results of our investigation suggested that the hospital cooling towers were the source of a community outbreak of LD. Increasing proximity to and frequency of exposure to the towers increased the risk of LD. New guidelines for cooling tower maintenance are needed. Knowing the location of cooling towers could facilitate maintenance inspections and outbreak investigations.

Impact of quality control on accuracy in enzyme immunoassay testing for human immunodeficiency virus type 1 antibodies.

OBJECTIVE: To assess use of quality control (QC) material, supplemental to internal kit controls (calibrators), as protection against errors in enzyme immunoassay testing for human immunodeficiency virus type 1 antibodies. DESIGN: From August 1994 to January 1996, enzyme immunoassay testing accuracy was assessed for laboratories participating in the Centers for Disease Control and Prevention Model Performance Evaluation Program that provided information regarding their use of QC material. Error rates were examined for human immunodeficiency virus type 1 antibody-negative, strongly positive, and weakly positive samples. RESULTS: The overall error rate with QC (2.20%) was significantly (P = .0023) lower than the error rate without QC (2.90%). With QC use there was a significant reduction in the relative risk of error for negative (P = .014) and weakly positive (P = .0067) samples. After multivariate analysis, use of QC lowered overall error rate by 29% (P = .0009). Laboratories not using QC were at increased risk of systematic error. Following the Clinical Laboratory Improvement Amendments of 1988 guidelines for QC material was relatively more protective against error than lower frequencies/number of levels. CONCLUSIONS: Using QC protected against errors in enzyme immunoassay testing for human immunodeficiency virus type 1 antibodies. Two levels of QC should be used with each run as mandated by the Clinical Laboratory Improvement Amendments of 1988.

Quantifying the bias associated with use of discrepant analysis.

Discrepant analysis is a widely used technique for estimating the performance parameters of a laboratory test. In discrepant analysis, each specimen is initially tested with the candidate test and a comparison method, and when the results of the two tests disagree, a confirmatory test is used to resolve the discrepancy. Discrepant analysis usually produces biased estimates. This report quantifies this bias and shows that it is usually positive, leading to overestimation of the performance parameters of a laboratory test. The direction and magnitude of the bias are predictably influenced by the analytical sensitivity and specificity of the candidate test, comparison method, and confirmatory test. The proportion of abnormal specimens tested also affects the magnitude of the bias, particularly the estimates of analytical sensitivity and positive predictive value when this proportion is low. Alternative approaches are suggested.

Incidence of community-acquired pneumonia requiring hospitalization. Results of a population-based active surveillance Study in Ohio. The Community-Based Pneumonia Incidence Study Group.

BACKGROUND: Pneumonia is the leading cause of death due to infectious diseases in the United States; however, the incidence of most infections causing community-acquired pneumonia in adults is not well defined. METHODS: We evaluated all adults, residing in 2 counties in Ohio, who were hospitalized in 1991 because of community-acquired pneumonia. Information about risk factors, symptoms, and outcome was collected through interview and medical chart review. Serum samples were collected from consenting individuals during the acute and convalescent phases, and specific etiologic diagnoses were assigned based on results of bacteriologic and immunologic tests. RESULTS: The incidence of community-acquired pneumonia requiring hospitalization in the study counties in 1991 was 266.8 per 100,000 population; the overall case-fatality rate was 8.8%. Pneumonia incidence was higher among blacks than whites (337.7/100,000 vs 253.9/ 100,000; P < .001), was higher among males than females (291.4 vs 244.8; P < .001), and increased with age (91.6/100,000 for persons aged < 45 years, 277.2/ 100,000 for persons aged 45-64 years, and 1012.3/ 100,000 for persons aged > or = 65 years; P < .001). Extrapolation from study incidence data showed the projected annual number of cases of community-acquired pneumonia requiring hospitalization in the United States to be 485,000. These data provide previously unavailable estimates of the annual number of cases that are due to Legionella species (8000-18,000), Mycoplasma pneumoniae (18,700-108,000), and Chlamydia pneumoniae (5890-49,700). CONCLUSIONS: These data provide information about the importance of community-acquired pneumonia and the relative and overall impact of specific causes of pneumonia. The study provides a basis for choosing optimal empiric pneumonia therapy, and allows interventions for prevention of pneumonia to be targeted at groups at greatest risk for serious illness and death.

Risk factors for domestic acquisition of legionnaires disease. Ohio legionnaires Disease Group.

BACKGROUND: Legionnaires disease is a common cause of adult pneumonia. Outbreaks of legionnaires disease have been well described, but little is known about sporadically occurring legionnaires disease, which accounts for most infections. Exposure to contaminated residential water sources is I plausible means of disease acquisition. METHODS: Employing a matched case-control study design in 15 hospitals in 2 Ohio counties, we prospectively enrolled 146 adults diagnosed as having nonepidemic, community-acquired legionnaires disease and compared each with 2 hospital-based control patients, matched for age, sex, and underlying illness category. An interview regarding potential exposures was followed by a home survey that included sampling residential sources for Legionella. Interview and home survey data were analyzed to estimate the risk of acquiring legionnaires disease associated with various exposures. RESULTS: Multivariate analysis showed that a nonmunicipal water supply (odds ratio [OR], 2.26; 95% confidence interval [CI], 1.17-4.37), recent residential plumbing repair (OR, 2.39; 95% CI, 1.10-5.18), and smoking (OR, 3.48; 95% CI, 2.09-5.79) were independent risk factors for legionnaires disease. Univariate analysis suggested that electric (vs gas) water heaters (OR, 1.97; 95% CI, 1.10-3.52), working more than 40 hours weekly (OR, 2.13; 95% CI, 1.12-4.07), and spending nights away from home before illness (OR, 1.68; 95% CI, 1.03-2.74) were additional possible risk factors. Lower chlorine concentrations in potable water and lower water heater temperatures were associated with residential Legionella colonization. CONCLUSIONS: A proportion of sporadic cases of legionnaires disease may be residentially acquired and are associated with domestic potable water and disruptions in residential plumbing systems. Potential strategies to reduce legionnaires disease risk include consistent chlorination of potable water, increasing water heater temperatures, and limiting exposure to aerosols after domestic plumbing repairs.

Risk factors for carriage of drug-resistant Streptococcus pneumoniae among children in Memphis, Tennessee.

OBJECTIVES: To determine risk factors for carriage of drug-resistant Streptococcus pneumoniae to understand better the factors promoting spread of these isolates. STUDY DESIGN: We obtained medical and demographic information and nasopharyngeal swab specimens from 216 children less than 6 years old with upper respiratory tract infections, seeking medical care at five Memphis, Tenn, study sites. We evaluated risk factors for carriage of penicillin-nonsusceptible S. pneumoniae (NSSP) among 100 children with S. pneumoniae isolates. Patterns of antimicrobial prescription were recorded for enrolled children. RESULTS: Independent risk factors for carriage of NSSP included an increased number of antimicrobial treatment courses during the previous 3 months and white race. Day care attendance approached statistical significance (p = 0.07). Most children with upper respiratory tract infection received a prescription for antimicrobial drugs. These prescriptions were more common for white children than for black children. CONCLUSIONS: Increased use of antimicrobial drugs enhances the risk of carriage of NSSP. This may contribute to the higher risk among white children of NSSP infection; however, after control for antimicrobial use, white children were still at an increased risk of infection with NSSP, possibly through greater exposure to resistant strains.

Outbreak of Legionnaires' disease among cruise ship passengers exposed to a contaminated whirlpool spa.

BACKGROUND: Outbreaks of travel-related Legionnaires' disease present a public-health challenge since rapid, sensitive, and specific diagnostic tests are not widely used and because detection of clusters of disease among travellers is difficult. We report an outbreak of Legionnaires' disease among cruise ship passengers that occurred in April, 1994, but that went unrecognised until July, 1994. METHODS: After rapid diagnosis of Legionnaires' disease in three passengers by urine antigen testing, we searched for additional cases of either confirmed (laboratory evidence of infection) or probable Legionnaires' disease (pneumonia of undetermined cause). A case-control study was conducted to compare exposures and activities on the ship and in ports of call between each case-passenger and two or three matched control-passengers. Water samples from the ship, from sites on Bermuda, and from the ship's water source in New York City were cultured for legionellae and examined with PCR. FINDINGS: 50 passengers with Legionnaires' disease (16 confirmed, 34 probable) were identified from nine cruises embarking between April 30 and July 9, 1994. Exposure to whirlpool spas was strongly associated with disease (odds ratio 16.2, 95% Cl 2.8-351:7); risk of acquiring Legionnaires' disease increased by 64% (95% Cl 12-140) for every hour spent in the spa water. Passengers spending time around the whirlpool spas, but not in the water, were also significantly more likely to have acquired infection. Legionella pneumophila serogroup 1 was isolated only from the sand filter in the ship's whirlpool spa. This isolate matched a clinical isolate from the respiratory secretions of a case-passenger as judged by monoclonal antibody subtyping and by arbitrarily primed PCR. INTERPRETATION: This investigation shows the benefit of obtaining a recent travel history, the usefulness or urine antigen testing for rapid diagnosis of legionella infection, and the need for improved surveillance for travel-related Legionnaires' disease. New strategies for whirlpool spa maintenance and decontamination may help to minimise transmission of legionellae from these aerosol-producing devices.

Clinical features that differentiate hantavirus pulmonary syndrome from three other acute respiratory illnesses.

To elucidate the early clinical characteristics of hantavirus pulmonary syndrome (HPS), we compared the clinical features of 24 cases of HPS with those of cases of bacteremic pneumococcal pneumonia (n = 30), influenza (n = 33), or unexplained adult respiratory distress syndrome (ARDS, n = 21). On admission, patients with HPS were less likely than outpatients with influenza to have reported sore throat (OR = 0.02, P < .01) and cough (OR = 0.1, P = .01) and were less likely than patients with pneumococcal pneumonia to have lobar infiltrates detected by chest roentgenography (OR = 0, P < .01). Multivariate discriminant analysis revealed that three clinical characteristics at admission (dizziness, nausea or vomiting, and absence of cough) and three initial laboratory abnormalities (low platelet count, low serum bicarbonate level, and elevated hematocrit level) served to identify all patients with HPS and to exclude HPS in at least 80% of patients with unexplained ARDS. These findings warrant further study and should facilitate the early recognition of patients with HPS, who may benefit from early critical-care intervention.

High prevalence of multidrug-resistant Streptococcus pneumoniae among children in a rural Kentucky community.

In 1992 drug-resistant Streptococcus pneumoniae was cultured with increasing frequency from aspirates of middle ear fluid from children with acute otitis media in a rural Kentucky community. To determine the prevalence of carriage of drug-resistant S. pneumoniae in the community, we obtained nasopharyngeal swabs from 158 (70%) of 227 children attending a child daycare center and from 82 children attending the county health center. S. pneumoniae was isolated from 126 children. Among 123 isolates tested 65 (53%) were penicillin-resistant, including 41 (33%) strains that were highly resistant; 61 (50%) were multidrug-resistant. Serotypes 19F, 6B, 23F and 6A comprised 89% of the penicillin-resistant isolates. Detection of a variety of serotypes and drug resistance patterns among nasopharyngeal isolates of S. pneumoniae suggests that multidrug-resistant pneumococcal strains are endemic in this community. Surveillance for drug-resistant pneumococci with the use of respiratory secretions obtained by nasopharyngeal swab may provide useful information on the prevalence of drug-resistant strains causing invasive disease and otitis media. Such information could be used to guide empiric therapy of pneumococcal infections.

Serotype distribution of Streptococcus pneumoniae infections among preschool children in the United States, 1978-1994: implications for development of a conjugate vaccine.

Conjugation of pneumococcal polysaccharide antigens to a protein carrier may improve protective immunity after vaccination of young children, an age group with high incidence of Streptococcus pneumoniae infection and poor immune responses to polysaccharide vaccines. To identify serotypes most commonly associated with infection in young children, pneumococcal isolates were serotyped from 3884 children < 6 years old (including 3007 < 2 years old) with pneumococcal bacteremia (n = 3169), meningitis (n = 401), or otitis media (n = 314). The isolates were submitted as part of a national surveillance during 1978-1994. Seven serotypes (14, 6B, 19F, 18C, 23F, 4, and 9V) accounted for 3045 isolates (78%). A conjugate pneumococcal vaccine protecting against these seven serotypes and serologically cross-reactive serotypes could potentially prevent 86% of bacteremia and 83% of meningitis but only 65% of otitis media cases. The proportion of isolates covered by such a vaccine increased from 78% to 87% during 1978-1994. Surveillance for pneumococcal serotypes causing infection is needed to detect shifts in serotype distribution over time.

Emergence of Buruli ulcer disease in the Daloa region of Cote d'Ivoire.

Recent reports have suggested increases in Buruli ulcer (BU), an infection caused by Mycobacterium ulcerans in west Africa. In 1991, we conducted surveillance for BU in a rural area of Cote d'Ivoire and identified 312 cases of active or healed ulceration. A case-control study was then performed to investigate risk factors for this infection. The rate of illness did not appear to differ between males and females (5.2% versus 7.5%; P = 0.11). The highest rate of illness was seen in the 10-14-year-old age group (143 cases per 1,000 population). New cases increased more than three-fold between 1987 and 1991, and local prevalence of BU was as high as 16.3%. Twenty-six percent of persons with healed ulcers had chronic functional disability. Participation in farming activities near the main river in the region was identified in the case-control study as a risk factor for infection (odds ratio [OR] for each 10-min decrease in walking distance between the fields and the river = 1.52, 95% confidence interval [CI] 1.01, 2.28, P = 0.046). Wearing long pants was protective (OR 0.20, 95% CI 0.06, 0.62, P < 0.005). We conclude that the incidence of BU is increasing rapidly in Cote d'Ivoire. Specific causes of this increase were not identified, but wearing protective clothing appeared to decrease the risk of disease.

Surveillance for Legionnaires' disease. Risk factors for morbidity and mortality.

BACKGROUND: To augment available information about the epidemiology of legionnaires' disease, we analyzed data reported to the passive surveillance system at the Centers for Disease Control and Prevention, Atlanta, Ga, from 1980 through 1989. METHODS: Risk of disease associated with specific demographic characteristics and health conditions was calculated by comparing the surveillance group with the US population. Risk of death was calculated using multivariate logistic regression models. RESULTS: A diagnosis of legionnaires' disease was confirmed on the basis of clinical and laboratory criteria for 3254 patients. Disease rates did not vary by year, but were higher in the northern states and during the summer. Legionella pneumophila, serogroup 1, constituted 71.5% of fully identified isolates. This study confirmed previously identified risk factors for legionnaires' disease. In addition, a markedly elevated risk was identified for persons with acquired immunodeficiency syndrome (rate ratio, 41.9; 95% confidence interval, 12.9, 71.0), or hematologic malignancy (rate ratio, 22.4; 95% confidence interval, 19.0, 25.9). Likelihood of death was increased in patients who were elderly or male; those with hospital-acquired infection, renal disease, malignancy, or immunosuppression; and those from whom L pneumophila, serogroup 6, was isolated. CONCLUSIONS: Infection with Legionella remains an important cause of disease and death in the United States. Diagnosis and treatment of legionnaires' disease should be targeted at patients at increased risk for illness and complications due to Legionella infection. Diagnostic tests for legionnaires' disease based on species other than L pneumophila, serogroup 1, should be developed and tested. Recommendations for prevention of legionnaires' disease should be focused on settings where there are persons at greatest risk for illness or serious outcome.

Pneumococcal polysaccharide vaccine efficacy. An evaluation of current recommendations.

OBJECTIVE: To determine pneumococcal polysaccharide vaccine efficacy in selected populations at risk for serious pneumococcal infection for whom vaccination is currently recommended and to assess duration of protection after vaccination. DESIGN: Vaccine efficacy was estimated using indirect cohort analysis to compare the proportion of pneumococcal infections caused by serotypes included in the vaccines of vaccinated and unvaccinated persons who were identified during 14 years of national surveillance. SETTING: Hospital laboratories in the United States that submitted pneumococcal isolates to the Centers for Disease Control and Prevention between May 1978 and April 1992. PARTICIPANTS: A total of 2837 persons older than 5 years who had pneumococcus isolated from blood or cerebrospinal fluid. RESULTS: Overall efficacy for preventing infection caused by serotypes included in the vaccine was 57% (95% confidence interval [CI], 45% to 66%). Efficacy among persons with diabetes mellitus was 84% (95% CI, 50% to 95%); with coronary vascular disease, 73% (95% CI, 23% to 90%); with congestive heart failure, 69% (95% CI, 17% to 88%); with chronic pulmonary diseases, 65% (95% CI, 26% to 83%); and with anatomic asplenia, 77% (95% CI, 14% to 95%). Efficacy was not documented for patients with alcoholism or cirrhosis, sickle cell disease, chronic renal failure, lymphoma, leukemia, or multiple myeloma, although sample sizes were small for these groups. Efficacy for immunocompetent persons older than 65 years was 75% (95% CI, 57% to 85%). Efficacy did not decline with increasing interval after vaccination: 5 to 8 years after vaccination it was 71% (95% CI, 24% to 89%), and 9 years or more after vaccination it was 80% (95% CI, 16% to 95%). CONCLUSIONS: Intensified efforts to improve pneumococcal vaccine coverage among certain populations for whom vaccination is currently recommended is indicated, but universal revaccination is not warranted at this time.