Margarette S. Kolczak

Mortality from invasive pneumococcal pneumonia in the era of antibiotic resistance, 1995-1997.

OBJECTIVES This study examined epidemiologic factors affecting mortality from pneumococcal pneumonia in 1995 through 1997. METHODS Persons residing in a surveillance area who had community-acquired pneumonia requiring hospitalization and Streptococcus pneumoniae isolated from a sterile site were included in the analysis. Factors affecting mortality were evaluated in univariate and multivariate analyses. The number of deaths from pneumococcal pneumonia requiring hospitalization in the United States in 1996 was estimated. RESULTS Of 5837 cases, 12% were fatal. Increased mortality was associated with older age, underlying disease. Asian race, and residence in Toronto/Peel, Ontario. When these factors were controlled for, increased mortality was not associated with resistance to penicillin or cefotaxime. However, when deaths during the first 4 hospital days were excluded, mortality was significantly associated with penicillin minimum inhibitory concentrations of 4.0 or higher and cefotaxime minimum inhibitory concentrations of 2.0 or higher. In 1996, about 7000 to 12,500 deaths occurred in the United States from pneumococcal pneumonia requiring hospitalization. CONCLUSIONS Older age and underlying disease remain the most important factors influencing death from pneumococcal pneumonia. Mortality was not elevated in most infections with beta-lactam-resistant pneumococci.

Active surveillance of vaccine safety : A system to detect early signs of adverse events

Background: There currently are no population-based systems in the United States to rapidly detect adverse events after newly introduced vaccines. To evaluate the feasibility of developing such systems, we used 5 years of data from 4 health maintenance organizations within the Centers for Disease Control and Prevention (CDC) Vaccine Safety Datalink. Methods: Within every year, each weeks vaccinated children were followed for 4 weeks, and rates of adverse events were compared with rates among children of similar ages before the introduction of the new vaccine. We assessed risks for intussusception after rotavirus vaccination and risks for fever, seizures, and other neurologic adverse events after the change from whole cell diphtheria-tetanus-pertussis (DTPw) to acellular DTP vaccine (DTPa). We used sequential probability ratio testing, adjusted for age, sex, calendar time, season, and HMO, and with a stopping value based on the probability of an adverse event under the null hypothesis and under a preset alternative hypothesis. Results: We detected an increase in intussusception after 2589 vaccine doses of rotavirus vaccine, about the same time initial reports of intussusception were made to the Vaccine Adverse Events Reporting System. Decreases in risk for fever, seizures, and other abnormal neurologic events became detectable within 12 weeks, 42 weeks, and 18 months, respectively, after the change from DTPw to DTPa. Conclusions: We conclude that it is feasible to develop systems for rapid and routine population-based assessments of new vaccine safety.

A Maximized Sequential Probability Ratio Test for Drug and Vaccine Safety Surveillance

Abstract Because of rare but serious adverse events, pharmaceutical drugs and vaccines are sometimes withdrawn from the market, either by a government agency such as the Food and Drug Administration (FDA) in the United States or by the manufacturing pharmaceutical company. In other cases, a drug may be generally safe but increase the risk for serious adverse events for certain subpopulations such as pregnant women or people with heart problems. Due to limited sample size and selected study populations, rare adverse events are often impossible to detect during phase 3 trials conducted before the drug is approved for general use. It is then important to conduct post-approval drug safety surveillance, using, for example, health insurance claims data. In such surveillance, the goal should be to detect serious adverse events as early as possible without too many false alarms, and it is then natural to use a continuous or near-continuous sequential test procedure that reevaluates the data on a daily or weekly basis. In this article, we first show that Walds classical sequential probability ratio test (SPRT) for continuous surveillance is very sensitive to the choice of relative risk required in the specification of the alternative hypothesis, making it difficult to use for drug and vaccine safety surveillance. We instead propose the use of a maximized sequential probability ratio test (MaxSPRT) based on a composite alternative hypothesis, which works well across a range of relative risks. We illustrate the use of this method on vaccine safety surveillance and compare it with the classical SPRT. A table of critical values for the MaxSPRT is provided, covering most parameter choices relevant for vaccine and drug safety surveillance. The critical values are based on exact numerical calculations. We also calculate the statistical power, the expected time until the null hypothesis is rejected, and the average length of surveillance.

Effect of Permethrin-Impregnated Nets on Exiting Behavior, Blood Feeding Success, and Time of Feeding of Malaria Mosquitoes (Diptera: Culicidae) in Western Kenya

Abstract The impact of permethrin-treated bednets on the feeding and house entering/exiting behavior of malaria vectors was assessed in two studies in western Kenya. In one study, matched pairs of houses were allocated randomly to receive bednets or no bednets. Exiting mosquitoes were collected in Colombian curtains hung around half of each house; indoor resting mosquitoes were collected by pyrethrum spray catches. The number of Anopheles gambiae Giles and An. arabiensis Patton estimated to have entered the houses was unaffected by the presence of bednets; Anopheles funestus Giles was less likely to enter a house if bednets were present. Anopheles gambiae and An. funestus were less likely to obtain a blood meal and significantly more likely to exit houses when bednets were present. No difference was detected in An. arabiensis rates of blood feeding and exiting. In a second experiment, hourly night biting collections were done on 13 nights during the rainy season to assess whether village-wide use of permethrin-treated bednets caused a shift in the time of biting of malaria vectors. A statistically significant shift was detected in the biting times of An. gambiae s.l., although the observed differences were small. No change was observed in the hourly distribution of An. funestus biting. Our study demonstrated that, at least in the short-term, bednets reduced human-vector contact and blood feeding success but did not lead to changes in the biting times of the malaria vectors in western Kenya.

Clinical outcomes of meningitis caused by Streptococcus pneumoniae in the era of antibiotic resistance.

Limited data are available on clinical outcomes of meningitis due to cefotaxime-nonsusceptible Streptococcus pneumoniae. We analyzed data from 109 cases of pneumococcal meningitis in Atlanta, Baltimore, and San Antonio, which were identified through population-based active surveillance from November 1994 to April 1996. Pneumococcal isolates from 9% of the cases were resistant to cefotaxime, and isolates from 11% had intermediate susceptibility. Children were more likely to have cephalosporin-nonsusceptible pneumococcal meningitis, but mortality was significantly higher among adults aged 18-64 years. Vancomycin was given upon admission to 29% of patients, and within 48 h of admission to 52%. Nonsusceptibility to cefotaxime was not associated with the following outcomes: increased mortality, prolonged length of hospital or intensive care unit (ICU) stay, requirement of intubation or oxygen, ICU care, discharge to another medical or long-term-care facility, or neurological deficit. Empirical use of vancomycin, current prevalence of drug-resistant S. pneumoniae, and degree of nonsusceptibility to cefotaxime may have influenced these findings.

Treatment History and Treatment Dose Are Important Determinants of Sulfadoxine-Pyrimethamine Efficacy in Children with Uncomplicated Malaria in Western Kenya

This study retrospectively studied amendable determinants of sulfadoxine-pyrimethamine (SP) efficacy involving 2869 treatments among 1072 Kenyan children <5 years old who had uncomplicated malaria. The dose was based on age: one-quarter tablet was given to infants <1 year old, one-half tablet was given to 1-3-year-old children, and a full tablet was given to 4-year-old children. Only 23.5% received the internationally recommended target dose of 25/1.25 mg of SP per kg of body weight. SP intake in the previous 15-35 days (adjusted relative risk, 1.67; 95% confidence interval, 1.35-2.07) and low SP dose (<27.5/1.375 mg/kg) (adjusted relative risk, 1.58; 95% confidence interval, 1.17-2.13) explained 38% of parasitological treatment failures by day 7. Patients with recent SP intake are likely to have recrudescent infections and may need close follow-up if treated with SP or alternative treatment. Applying our weight-for-age data to 31 existing age-based SP dose recommendations predicted that 22 of them would result in underdosing of >25% of children <5 years. Many age-based dose recommendations need urgent revision, because SP is increasingly used as first-line treatment in sub-Saharan Africa.

Increased efficacy of sulfadoxine-pyrimethamine in the treatment of uncomplicated falciparum malaria among children with sickle cell trait in western Kenya.

The role of the sickle cell hemoglobin type as a determinant of treatment outcome with sulfadoxine-pyrimethamine was retrospectively studied in young children with uncomplicated falciparum malaria who lived in an area with intense perennial malaria transmission. Between 1993 and 1997, 2795 treatments involving 813 children were monitored. Sickle cell trait (HbAS) was present in 17.7% of the children. Two-and-a-half percent of the children experienced early clinical treatment failure by day 2-3, and 17.3% of the children were parasitemic on day 7. Treatments in HbAS children were less likely than those in HbAA children to result in persistence of parasitemia by day 3 (relative risk [RR], 0.66; 95% confidence interval [CI], 0.47-0.93; P=.02) or in parasitologic treatment failure on day 7 (RR, 0.51; 95% CI, 0.36-0.71; P<.0001). These results suggest that the HbAS phenotype should be included among factors that determine sulfadoxine-pyrimethamine treatment outcome.

Variation in hepatitis B immunization coverage rates associated with provider practices after the temporary suspension of the birth dose

BackgroundIn 1999, the American Academy of Pediatrics and U.S. Public Health Service recommended suspending the birth dose of hepatitis B vaccine due to concerns about potential mercury exposure. A previous report found that overall national hepatitis B vaccination coverage rates decreased in association with the suspension. It is unknown whether this underimmunization occurred uniformly or was associated with how providers changed their practices for the timing of hepatitis B vaccine doses. We evaluate the impact of the birth dose suspension on underimmunization for the hepatitis B vaccine series among 24-month-olds in five large provider groups and describe provider practices potentially associated with underimmunization following the suspension.MethodsRetrospective cohort study of children enrolled in five large provider groups in the United States (A-E). Logistic regression was used to evaluate the association between the birth dose suspension and a childs probability of being underimmunized at 24 months for the hepatitis B vaccine series.ResultsPrior to July 1999, the percent of children who received a hepatitis B vaccination at birth varied widely (3% to 90%) across the five provider groups. After the national recommendation to suspend the hepatitis B birth dose, the percent of children who received a hepatitis B vaccination at birth decreased in all provider groups, and this trend persisted after the policy was reversed. The most substantial decreases were observed in the two provider groups that shifted the first hepatitis B dose from birth to 5–6 months of age. Accounting for temporal trend, children in these two provider groups were significantly more likely to be underimmunized for the hepatitis B series at 24 months of age if they were in the birth dose suspension cohort compared with baseline (Group D OR 2.7, 95% CI 1.7 – 4.4; Group E OR 3.1, 95% CI 2.3 – 4.2). This represented 6% more children in Group D and 9% more children in Group E who were underimmunized in the suspension cohort compared with baseline. Children in the reversal cohort in these groups remained significantly more likely to be underimmunized compared with baseline. In contrast, in a third provider group where the typical timing of the third dose was unchanged and in two other provider groups whose hepatitis B vaccination schedules were unaffected by the birth dose suspension, hepatitis B vaccination coverage either was maintained or improved.ConclusionWhen the hepatitis B birth dose was suspended, provider groups that moved the first dose of vaccination to 5–6 months of age or later had decreases in hepatitis B vaccine coverage at 24 months. These findings suggest that as vaccine policy changes occur, providers could attempt to minimize underimmunization by adopting vaccination schedules that minimize delays in the recommended timing of vaccine doses.

Subclinical Plasmodium falciparum Infection and HIV-1 Viral Load

To the Editor: Studies indicate that Plasmodium falciparum infection increases HIV replication in adults (1,2). Although malaria-related illness and death are more common in children, and HIV-1 generally progresses faster in children than in adults (3,4), to our knowledge the effect of intermittent malaria on HIV-1 viral load has not been directly explored in children. To investigate this issue, we monitored HIV-positive infants from a 1996–2001 birth cohort study in Kisumu, Kenya, a P. falciparum–holoendemic area. Study design and methods have been described elsewhere (5,6). Twenty-four children that were perinatally infected with HIV were included in this substudy. During monthly visits during the child’s first 2 years of life, malaria and HIV incidence were recorded (5,6). Both children with malaria-positive blood smears and those with fever but no smear result available were treated with sulfadoxine-pyrimethamine according to national guidelines. At the time of this substudy, none of the study participants were taking antiretroviral drugs. HIV and malaria diagnoses were determined by using standard methods (5–7). To reduce the chance of including infants infected through breast-feeding, perinatal infection was defined as >2 consecutive HIV-positive tests, with the first positive PCR results by 4 months of age (7). The so-called baseline viral load was the premalaria value measured 1 month before the first observation in the analysis. To be included in the analysis, follow-up visits had to have data available on the current and previous months’ viral load and malaria status and occur at roughly monthly intervals at >4 months of age. Malaria parasites were found at 53 of 146 visits in the month before viral load measurement, although at 89% of visits in which children were malaria-positive, the children’s samples had <1,000 parasites/μL, and in only 13% of visits in which children had parasitemia did they also have fever (8). Median number of observations per child was 7 (range 2–18). No significant demographic or clinical differences were found between HIV-positive children in this substudy and those enrolled in the full cohort (data not shown). Clinical and demographic variables were evaluated in univariate repeated measures analysis to determine associations with log-transformed HIV-1 viral load. Age and baseline viral load were strong predictors of current load (Table). Although not statistically significant, clearing the previous month’s malaria infection was associated with a drop in viral load (Table, p = 0.09). It was not possible to distinguish between the effects of treatment versus malaria clearance because 87% of malaria infections were treated with antimalarial drugs. However, viral load increased in those incorrectly treated for malaria presumptively (Table). Table Associations with log HIV-1 viral load in infants* After adjusting for age and baseline viral load, we assessed log10 HIV viral load in relation to malaria clearance, persistence, absence, or new infection using a repeated measures model with autoregressive covariance structure. No differences were statistically significant, although clearing an infection versus no malaria had a 0.22 log viral load decrease (Table, p = 0.10). When 15 malaria episodes with malaria-free visits 1 month before and after the episode were compared, mean difference (signed-rank test) in viral load “before” and “after” malaria was not significant. Our findings suggest that low-density malaria infection may not dramatically affect plasma HIV-1 levels in infants. This finding is similar to results of studies of perinatally HIV-infected children in which, although viral loads were unavailable, number of malaria episodes did not significantly affect development of AIDS-related symptoms (9,10). While clinical malaria leads to at least short-term HIV viral load increases in adults (1,2), the effect of subclinical malaria is unclear, and even less is known about coinfection in children. Children usually have higher baseline viral loads than adults; thus, the relative effect of malaria on viral load may not be as great. To reduce the impact of passively transferred maternal antibodies, analyses were done on visits after the child was 4 months old. However, lack of fully acquired antimalarial immunity may have led to different HIV/malaria interactions than seen in adults. Viral load increased in infants that were incorrectly treated presumptively (due to fever) for malaria (Table). Most of these children were found to have other infections. Fever in malaria-endemic areas is often assumed to be malaria-related, but delay in treatment of nonmalarial infections may be harmful in HIV-infected children Our assessment was limited in size and duration. Furthermore, in attempting to provide optimal patient care through conducting monthly surveillance and encouraging mothers to bring children in during febrile episodes, ability to assess the effect of high-density malaria was diminished because parasitemia levels never reached clinically significant levels. Finally, because malaria was diagnosed by microscopy, rather than PCR, some subclinical malaria infections may have been misclassified as malaria-negative. Although we found no evidence of an association between subclinical, low-density malaria and infant HIV-1 viral load, the consequences of high-density or clinical malaria need to be explored. If clinical malaria in infants increases HIV-1 viral load as it does in adults (1,2), our study underscores dual benefits of malaria treatment in the context of HIV: 1) keeping malaria in check, and 2) preventing an increase in HIV viral load. Ethical issues prevent prospective studies to assess the impact of coinfection early in life, but alternatives include using animal models or stored specimens.

Multiple vaccinations and the risk of medically attended fever.

Recent increases in the number of vaccinations recommended for infants have triggered concerns about the safety of multiple vaccinations. This study evaluated rates of medically attended fever after infant vaccination using computerized data from 1991 to 2000 from two large U.S. provider groups. The rate of medically attended fever within 7 days after vaccination was low (6.4 per 1000 vaccination visits) and did not increase during the decade. Higher rates of fever occurred during periods when a third dose of oral polio vaccine was used (1994-1995) and when a now-discontinued oral rotavirus vaccine was used (1998-1999). These findings offer reassurance that the multiple vaccinations introduced during the decade studied were not associated with increases in medically attended fever.