Epidemiologic programs for computers and calculators. Exact binomial confidence intervals for the relative risk in follow-up studies with sparsely stratified incidence density data

Estimation of Relative Vaccine Effectiveness in Influenza: A Systematic Review of Methodology

BACKGROUND

When new vaccine components or platforms are developed, they will typically need to demonstrate noninferiority or superiority over existing products, resulting in the assessment of relative vaccine effectiveness (rVE). This review aims to identify how rVE evaluation is being performed in studies of influenza to inform a more standardized approach.

METHODS

We conducted a systematic search on PubMed, Google Scholar, and Web of Science for studies reporting rVE comparing vaccine components, dose, or vaccination schedules. We screened titles, abstracts, full texts, and references to identify relevant articles. We extracted information on the study design, relative comparison made, and the definition and statistical approach used to estimate rVE in each study.

RESULTS

We identified 63 articles assessing rVE in influenza virus. Studies compared multiple vaccine components (n = 38), two or more doses of the same vaccine (n = 17), or vaccination timing or history (n = 9). One study compared a range of vaccine components and doses. Nearly two-thirds of all studies controlled for age, and nearly half for comorbidities, region, and sex. Assessment of 12 studies presenting both absolute and relative effect estimates suggested proportionality in the effects, resulting in implications for the interpretation of rVE effects.

CONCLUSIONS

Approaches to rVE evaluation in practice is highly varied, with improvements in reporting required in many cases. Extensive consideration of methodologic issues relating to rVE is needed, including the stability of estimates and the impact of confounding structure on the validity of rVE estimates.

Controversy and debate on dengue vaccine series-paper 1: review of a licensed dengue vaccine: inappropriate subgroup analyses and selective reporting may cause harm in mass vaccination programs

Severe life-threatening dengue fever usually occurs when a child is infected by dengue virus a second time. This is caused by a phenomenon called antibody-dependent enhancement (ADE). Since dengue vaccines can mimic a first infection in seronegative children (those with no previous infection), a natural infection later in life could lead to severe disease. The possibility that dengue vaccines can cause severe dengue through ADE has led to serious concern regarding the safety of mass vaccination programs. A published meta-analysis addressed this safety issue for a new vaccine against dengue fever-Dengvaxia. The trials in this meta-analysis have been used to campaign for mass vaccination programs in developing countries. We discuss the results of this paper and point out problems in the analyses. Reporting the findings in an Asian trial (CYD14), the authors show a sevenfold rise in one outcome-hospitalization for dengue fever in children <5 years old. However, they fail to point out two signals of harm for another outcome-hospitalization for severe dengue fever (as confirmed by an independent data monitoring committee): 1. In children younger than 9 years, the relative risk was 8.5 (95% confidence interval [CI]: 0.5, 146.8), and 2. In the overall study group, the relative risk was 5.5 (95% CI: 0.9, 33). The authors conduct a subgroup analysis to support claims that the vaccine is probably safe among children aged 9 years or more. This subgroup analysis has limited credibility because: (1) it was a post hoc analysis; (2) it was one of a large number of subgroup analyses; (3) the test of interaction was not reported, but was insignificant (P = 0.14); and (4) there is no biological basis for a threshold age of 9 years. The more likely explanation for the higher risk in younger children is ADE, that is, more frequent seronegativity, rather than age itself. The selective reporting and inappropriate subgroup claims mask the potential harm of dengue mass vaccination programs. Countries planning public use of the vaccine must conduct diligent postmarketing surveillance, secure informed consent from parents of potential recipients, and closely monitor the results of ongoing long-term follow-up of clinical trial participants.

Long-term persistence of zoster vaccine efficacy

BACKGROUND

The Shingles Prevention Study (SPS) demonstrated zoster vaccine efficacy through 4 years postvaccination. A Short-Term Persistence Substudy (STPS) demonstrated persistence of vaccine efficacy for at least 5 years. A Long-Term Persistence Substudy (LTPS) was undertaken to further assess vaccine efficacy in SPS vaccine recipients followed for up to 11 years postvaccination. Study outcomes were assessed for the entire LTPS period and for each year from 7 to 11 years postvaccination.

METHODS

Surveillance, case determination, and follow-up were comparable to those in SPS and STPS. Because SPS placebo recipients were offered zoster vaccine before the LTPS began, there were no unvaccinated controls. Instead, SPS and STPS placebo results were used to model reference placebo groups.

RESULTS

The LTPS enrolled 6867 SPS vaccine recipients. Compared to SPS, estimated vaccine efficacy in LTPS decreased from 61.1% to 37.3% for the herpes zoster (HZ) burden of illness (BOI), from 66.5% to 35.4% for incidence of postherpetic neuralgia, and from 51.3% to 21.1% for incidence of HZ, and declined for all 3 outcome measures from 7 through 11 years postvaccination. Vaccine efficacy for the HZ BOI was significantly greater than zero through year 10 postvaccination, whereas vaccine efficacy for incidence of HZ was significantly greater than zero only through year 8.

CONCLUSIONS

Estimates of vaccine efficacy decreased over time in the LTPS population compared with modeled control estimates. Statistically significant vaccine efficacy for HZ BOI persisted into year 10 postvaccination, whereas statistically significant vaccine efficacy for incidence of HZ persisted only through year 8.

Persistence of the efficacy of zoster vaccine in the shingles prevention study and the short-term persistence substudy

BACKGROUND

The Shingles Prevention Study (SPS; Department of Veterans Affairs Cooperative Study 403) demonstrated that zoster vaccine was efficacious through 4 years after vaccination. The Short-Term Persistence Substudy (STPS) was initiated after the SPS to further assess the persistence of vaccine efficacy.

METHODS

The STPS re-enrolled 7320 vaccine and 6950 placebo recipients from the 38 546-subject SPS population. Methods of surveillance, case determination, and follow-up were analogous to those in the SPS. Vaccine efficacy for herpes zoster (HZ) burden of illness, incidence of postherpetic neuralgia (PHN), and incidence of HZ were assessed for the STPS population, for the combined SPS and STPS populations, and for each year through year 7 after vaccination.

RESULTS

In the STPS as compared to the SPS, vaccine efficacy for HZ burden of illness decreased from 61.1% to 50.1%, vaccine efficacy for the incidence of PHN decreased from 66.5% to 60.1%, and vaccine efficacy for the incidence of HZ decreased from 51.3% to 39.6%, although the differences were not statistically significant. Analysis of vaccine efficacy in each year after vaccination for all 3 outcomes showed a decrease in vaccine efficacy after year 1, with a further decline thereafter. Vaccine efficacy was statistically significant for the incidence of HZ and the HZ burden of illness through year 5.

CONCLUSIONS

Vaccine efficacy for each study outcome was lower in the STPS than in the SPS. There is evidence of the persistence of vaccine efficacy through year 5 after vaccination but, vaccine efficacy is uncertain beyond that point.

A randomized controlled trial of motavizumab versus palivizumab for the prophylaxis of serious respiratory syncytial virus disease in children with hemodynamically significant congenital heart disease

Children with hemodynamically significant congenital heart disease (CHD) are at risk for serious respiratory syncytial virus (RSV) disease. This study was designed to assess the safety and tolerability of motavizumab versus palivizumab in children with CHD and was not powered for efficacy. Patients (n = 1236) aged ≤24 mo were randomized to receive five monthly doses (15 mg/kg) of motavizumab or palivizumab during the RSV season. Adverse events (AEs) and serious AEs (SAEs) were recorded through 30 d after the last dose. RSV hospitalizations and RSV outpatient medically attended lower respiratory tract infections (MALRI; season 2) were summarized. Approximately 93 and 50% of patients reported an AE or SAE, respectively. Skin events occurred in 19.3% of motavizumab recipients and 16.2% of palivizumab recipients. Rates of hospitalizations and RSV MALRI were similar between treatment groups [relative risk (RR): 0.75; 95% CI, 0.34-1.59 and RR: 0.49; 95% CI, 0.10-1.99, respectively; both p > 0.05]. Motavizumab and palivizumab had similar safety profiles in children with hemodynamically significantly CHD; with the exception of skin events which were increased in motavizumab recipients. Safety and efficacy were consistent with another study comparing motavizumab with palivizumab in premature infants without CHD.

Genetic sequence analysis of influenza viruses and illness severity in ill children previously vaccinated with live attenuated or inactivated influenza vaccine

In a large comparative study in 2004-2005, children aged 6-59 months vaccinated with live attenuated influenza vaccine (LAIV) experienced 55% fewer cases of culture-confirmed influenza illness compared with trivalent inactivated influenza vaccine (TIV) recipients. To better understand the characteristics of the breakthrough influenza illnesses, we analyzed the HA1 genetic sequence for all available samples and examined disease severity by strain and treatment group. All 48 A/H1N1 viruses were well-matched to the vaccine, whereas all 276 A/H3N2 viruses and 349 (96%) influenza B viruses were mismatched to the vaccine. The incidence of fever or lower respiratory illness did not differ by strain; however, LAIV recipients had less febrile disease and fewer lower respiratory illnesses than TIV recipients. Viruses of each influenza B lineage caused more illnesses than A/H1N1 viruses; strategies to enhance protection against multiple influenza B lineages should be pursued.

Motavizumab for prophylaxis of respiratory syncytial virus in high-risk children: a noninferiority trial

OBJECTIVE

Palivizumab reduces respiratory syncytial virus (RSV) hospitalization in children at high risk by approximately 50% compared with placebo. We compared the efficacy and safety of motavizumab, an investigational monoclonal antibody with enhanced anti-RSV activity in preclinical studies, with palivizumab.

METHODS

This randomized, double-blind, multinational, phase 3, noninferiority trial assessed safety and RSV hospitalization in 6635 preterm infants aged <or=6 months at enrollment or children aged <or=24 months with chronic lung disease of prematurity who received 15 mg/kg palivizumab or motavizumab monthly. Secondary end points included outpatient medically attended lower respiratory tract infections (MALRIs), RSV-specific LRIs, otitis media, antibiotic use, development of antimotavizumab antibodies, and motavizumab serum concentrations.

RESULTS

Motavizumab recipients had a 26% relative reduction in RSV hospitalization compared with palivizumab recipients, achieving noninferiority. Motavizumab was superior to palivizumab for reduction of RSV-specific outpatient MALRIs (50% relative reduction). Overall, adverse events (AEs) were not significantly different between groups. Cutaneous events were reported in 2 percentage points more motavizumab recipients (7.2% vs 5.1%); most were mild, but 0.3% resulted in dosing discontinuation. Antidrug antibodies (ADA) were detected in 1.8% of motavizumab recipients. Patients with anti-drug antibody reported 6 RSV events and 17 cutaneous events.

CONCLUSIONS

Children receiving prophylaxis with motavizumab or palivizumab had low rates of RSV hospitalization; motavizumab recipients experienced 50% fewer RSV MALRIs than palivizumab recipients. AEs were similar in both groups, although cutaneous AEs were higher for motavizumab recipients. Motavizumab may offer an improved alternative in prophylaxis for serious RSV disease in infants and children at high risk.

A vaccine to prevent herpes zoster and postherpetic neuralgia in older adults

BACKGROUND

The incidence and severity of herpes zoster and postherpetic neuralgia increase with age in association with a progressive decline in cell-mediated immunity to varicella-zoster virus (VZV). We tested the hypothesis that vaccination against VZV would decrease the incidence, severity, or both of herpes zoster and postherpetic neuralgia among older adults.

METHODS

We enrolled 38,546 adults 60 years of age or older in a randomized, double-blind, placebo-controlled trial of an investigational live attenuated Oka/Merck VZV vaccine ("zoster vaccine"). Herpes zoster was diagnosed according to clinical and laboratory criteria. The pain and discomfort associated with herpes zoster were measured repeatedly for six months. The primary end point was the burden of illness due to herpes zoster, a measure affected by the incidence, severity, and duration of the associated pain and discomfort. The secondary end point was the incidence of postherpetic neuralgia.

RESULTS

More than 95 percent of the subjects continued in the study to its completion, with a median of 3.12 years of surveillance for herpes zoster. A total of 957 confirmed cases of herpes zoster (315 among vaccine recipients and 642 among placebo recipients) and 107 cases of postherpetic neuralgia (27 among vaccine recipients and 80 among placebo recipients) were included in the efficacy analysis. The use of the zoster vaccine reduced the burden of illness due to herpes zoster by 61.1 percent (P<0.001), reduced the incidence of postherpetic neuralgia by 66.5 percent (P<0.001), and reduced the incidence of herpes zoster by 51.3 percent (P<0.001). Reactions at the injection site were more frequent among vaccine recipients but were generally mild.

CONCLUSIONS

The zoster vaccine markedly reduced morbidity from herpes zoster and postherpetic neuralgia among older adults.

Cardiovascular events associated with rofecoxib in a colorectal adenoma chemoprevention trial

BACKGROUND

Selective inhibition of cyclooxygenase-2 (COX-2) may be associated with an increased risk of thrombotic events, but only limited long-term data have been available for analysis. We report on the cardiovascular outcomes associated with the use of the selective COX-2 inhibitor rofecoxib in a long-term, multicenter, randomized, placebo-controlled, double-blind trial designed to determine the effect of three years of treatment with rofecoxib on the risk of recurrent neoplastic polyps of the large bowel in patients with a history of colorectal adenomas.

METHODS

A total of 2586 patients with a history of colorectal adenomas underwent randomization: 1287 were assigned to receive 25 mg of rofecoxib daily, and 1299 to receive placebo. All investigator-reported serious adverse events that represented potential thrombotic cardiovascular events were adjudicated in a blinded fashion by an external committee.

RESULTS

A total of 46 patients in the rofecoxib group had a confirmed thrombotic event during 3059 patient-years of follow-up (1.50 events per 100 patient-years), as compared with 26 patients in the placebo group during 3327 patient-years of follow-up (0.78 event per 100 patient-years); the corresponding relative risk was 1.92 (95 percent confidence interval, 1.19 to 3.11; P=0.008). The increased relative risk became apparent after 18 months of treatment; during the first 18 months, the event rates were similar in the two groups. The results primarily reflect a greater number of myocardial infarctions and ischemic cerebrovascular events in the rofecoxib group. There was earlier separation (at approximately five months) between groups in the incidence of nonadjudicated investigator-reported congestive heart failure, pulmonary edema, or cardiac failure (hazard ratio for the comparison of the rofecoxib group with the placebo group, 4.61; 95 percent confidence interval, 1.50 to 18.83). Overall and cardiovascular mortality was similar in the two groups.

CONCLUSIONS

Among patients with a history of colorectal adenomas, the use of rofecoxib was associated with an increased cardiovascular risk.

Safety of cold-adapted live attenuated influenza vaccine in a large cohort of children and adolescents

OBJECTIVE

To determine the safety of cold-adapted trivalent intranasal influenza virus vaccine (CAIV) in children and adolescents.

STUDY DESIGN

A randomized, double blind, placebo-controlled safety trial in healthy children age 12 months to 17 years given CAIV (FluMist; MedImmune Vaccines, Inc.) or placebo (randomization, 2:1). Children <9 years of age received a second dose of CAIV or placebo 28 to 42 days after the first dose. Enrolled children were then followed for 42 days after each vaccination for all medically attended events. Prespecified outcomes included 4 prespecified diagnostic groups and 170 observed individual diagnostic categories. The relative risk and the 2-sided 90% confidence interval were calculated for each diagnostic group and individual category by clinical setting, dose and age. More than 1500 relative risk analyses were performed.

RESULTS

A total of 9689 evaluable children were enrolled in the study. Of the 4 prespecified diagnostic categories (acute respiratory tract events, systemic bacterial infection, acute gastrointestinal tract events and rare events potentially associated with wild-type influenza), none was associated with vaccine. Of the biologically plausible individual diagnostic categories, 3, acute gastrointestinal events, acute respiratory events and abdominal pain, had different analyses that demonstrated increased and decreased relative risks, making their association with the vaccine unlikely. For reactive airway disease a significant increased relative risk was observed in children 18 to 35 months of age with a relative risk of 4.06 (90% confidence interval, 1.29 to 17.86) in this age group. The individual diagnostic categories of upper respiratory infection, musculoskeletal pain, otitis media with effusion and adenitis/adenopathy had at least one analysis that achieved a significant increased risk ratio. All of these events were infrequent.

CONCLUSION

CAIV was generally safe in children and adolescents. The observation of an increased risk of asthma/reactive airway disease in children <36 months of age is of potential concern. Further studies are planned to evaluate the risk of asthma/reactive airway disease after vaccine.

Post-marketing evaluation of the short term safety of COMVAX®

We report an evaluation of the short-term safety of a pediatric bivalent combination vaccine containing RECOMBIVAXHB and Liquid PedvaxHIB, COMVAX. Safety was assessed through identification of medical utilization; potential adverse events were identified through computerized clinical databases for deaths, hospitalizations, emergency room visits, and outpatient clinic visits. We calculated relative risks whenever there was at least one diagnosis-specific event in the risk period following vaccination and compared the rates in specific time windows following vaccination with rates at 31-60 days following vaccination and also with rates in a historical cohort of children. A total of 27,802 doses of COMVAX were administered, with 127 separate adverse event codes with statistically significant elevated risks, and 66 codes with significantly decreased risks. Most potentially serious diagnoses appeared in four major categories: "Respiratory Events"; "Gastroenteritis"; "Adverse Effect of Medicinal and Biological Substance, NOS"; and "Fever". There was no consistent pattern to indicate increased risks for serious respiratory or gastrointestinal illness. For fever, most of the findings appeared to be explained by changes in data collection or by concomitant vaccination with M-M-R(-)II. There was an increased risk for fever hospitalizations following shot 1. The total number of children hospitalized with fever was seven out of 12,468 children; all recovered fully. In this study population of 27,802 vaccine recipients, COMVAX appeared to have a favorable safety profile.

Incidence of myocardial infarction in randomized clinical trials of protease inhibitor-based antiretroviral therapy: an analysis of four different protease inhibitors

Protease inhibitor (PI) therapy for patients infected with the human immunodeficiency virus has been associated with lipid disorders and insulin resistance. We compared the incidence of myocardial infarction (MI) among participants receiving treatment with PIs with or without nucleoside reverse transcriptase inhibitors (nRTIs) to nRTI therapy alone in 30 phase II/III double-blind, randomized studies conducted before 1999 for the first 4 PI drugs. In most trials included in this analysis, participants could receive combination therapy with a PI plus nRTIs in open-label extensions after the blinded phase concluded. Person-years (PY) of follow-up were calculated from treatment initiation to the diagnosis of MI, or to the end of the randomized phases for nRTI-only therapy or to the conclusion of the studies for PI-containing regimens. Separate analyses were conducted for the randomized and the randomized-plus-extension phases. Among 10,986 participants, 7951 (72%) received PI drugs at some point for an average duration of 12 months. There were 10 MIs (1.31/1000 PY) in the randomized phases and 19 MIs (1.63/1000 PY) in the randomized-plus-extension phases. The overall stratified relative risk of MI for PI-containing (1.82 MI/1000 PY) versus nRTI-only (1.05 MI/1000 PY) regimens of 1.69 was not significantly increased (95% confidence interval [CI], 0.54 to 7.48). The absolute difference in MI risk was +0.77 (95% CI, -0.71 to +2.26) MIs/1000 PY. Compared with NRTI-only therapy, patients receiving PI-containing regimens for an average of 1 year did not have significantly more MIs, but the upper bound of the 95% CI indicates there may be up to 2.3 additional MIs per 1000 PY. Although studies with a longer duration of PI therapy are in progress to assess whether a later increase in MI incidence occurs, our analysis did not demonstrate a dramatic increase in MI risk during the first year of PI therapy.

Proving non-inferiority or equivalence of two treatments with dichotomous endpoints using exact methods

Since the early work of RA Fisher, exact methods have been recognized as important tools in data analysis because they provide valid statistical inference even with small sample sizes, or with sparse or skewed data. With the recent advance of computational power and the availability of commercial software packages, exact methods have gained substantial popularity over the past two decades. However, most of these exact methods have been devoted to testing classical null hypotheses of no differences, and until recently little was known about exact methods dealing with non-inferiority or equivalence hypotheses. The presence of nuisance parameters in testing non-inferiority/equivalence hypotheses presents a special challenge for exact methods because of the intense computational requirement. In this paper, we review exact methods available for proving non-inferiority or equivalence of two treatments with a dichotomous endpoint. First, we present the general methodology for conducting exact tests for non-inferiority or equivalence; we then discuss several unconditional and conditional methods available for constructing hypothesis tests and confidence intervals based on three commonly used measures, namely, the difference, relative risk, and odds ratio of two independent proportions or rates. Finally, we illustrate with several examples the application of these exact methods in analysing and planning non-inferiority or equivalence trials.

Indinavir did not increase the short-term risk of adverse cardiovascular events relative to nucleoside reverse transcriptase inhibitor therapy in four phase III clinical trials

A retrospective person-time analysis of the randomized and non-randomized extension phases of four phase III trials was performed to assess the incidence of adverse cardiovascular events in 2680 HIV-infected patients receiving indinavir or nucleoside reverse transcriptase inhibitor therapy, or both. The observed rate of cardiovascular events was not increased in patients receiving indinavir-based regimens compared with therapy without a protease inhibitor. Extrapolation of these findings is limited by the brief length of therapy and the small number of cases.

Preventability of vehicle-related fatalities

Several different methods have been used to evaluate the survivability of traffic injuries. Previously published methods were reviewed to develop a specific method for vehicle-related fatalities. All fatally injured victims (n = 474) of vehicle-related crashes in the four northern-most counties of Sweden during a 5-year period were studied. Almost half (48%) of the victims had non-survivable injuries. The remaining cases were classified into different injury groups according to the Injury Severity Score (ISS) including 56 (12%) with an ISS < 25, 150 (32%) with ISS 25-49 and 42 (9%) with ISS 50-74. The median distance from the scene to the nearest hospital was 49 km. In the cases with ISS < 50 the medical care was further analyzed. In close to half of these cases, the victim did not receive optimal care in time with transportation time being of major importance. The absence of first aid was judged to have contributed to the death in 4% of the cases.

Pedestrian fatalities and alcohol

Since the share of alcohol-positive victims has been reported not to show the same reduction in pedestrian fatalities as in other 'road users', this study analysed pedestrian fatalities with special emphasis on the presence of alcohol. Autopsied pedestrian fatalities (286 victims) in northern Sweden from 1977 to 1995 were investigated, using autopsy and police reports. Blood alcohol was detected in 19% of the fatalities, with a median concentration of 1.6 g/l. The proportion of test-positive victims did not change significantly during the period of study. Males more often tested positive for alcohol than females (24 vs. 11%). Further, the test-positive victims had a lower mean age (49 vs. 59 years) and the proportion of test-positive victims was higher during weekends and nights. Pedestrian fatalities testing positive for alcohol differ from those testing negative in several respects. The public should be enlightened of the fact that drinking is a risk not only among users of motor vehicles, but also among pedestrians. Separation of pedestrians from motor vehicles by, e.g. fences, is of greatest importance.

Postmarketing evaluation of the safety and effectiveness of varicella vaccine

BACKGROUND

The Oka strain of live attenuated varicella virus was licensed for use in healthy children in the United States in March, 1995. We report a postmarketing evaluation of the short term safety of this vaccine within Kaiser Permanente.

METHODS

After licensure varicella vaccination was introduced into the preventive care program of the Northern California Kaiser Permanente Medical Care Program. Potential adverse events after vaccination with varicella vaccine were identified from automated clinical databases of hospitalizations, emergency room visits and clinic visits. Deaths were identified from automated clinical databases at Kaiser as well as from the State death records for California. To evaluate safety, rates of diagnosis-specific events in the risk periods were compared with the rates of such diagnosis-specific events in two self control and one historical control period.

RESULTS

During the study period of April 1, 1995, to December 31, 1996, a total of 89753 adults and children received varicella vaccine. A total of 3200 relative risks were calculated, and of these 5 hospital diagnostic categories, 9 emergency visit diagnostic categories and 30 outpatient diagnostic categories demonstrated at least 1 relative risk with a P value of <0.05 in 1 or more age groups and in comparisons with 1 control period or more. The p value for these tests was not adjusted for multiple comparisons. Of these categories 14 demonstrated an increased risk either in more than 1 age group or against more than 1 comparison group. These categories included elective procedures, febrile seizure, febrile illness, well child, acute gastroenteritis, varicella, congenital anomaly, "rule out sepsis," trauma, viral syndrome, apnea, back pain, congenital valvular heart disease and vision evaluation for glasses. Of these the outcomes of elective procedure, congenital anomaly, congenital valvular heart disease, well child and vision evaluation for glasses were judged not to have a biologically plausible association with vaccination. A second diagnostic grouping included febrile illness, viral illness, febrile seizure and "rule out sepsis." In an analysis of these events which adjusted for the concomitant administration of M-M-R(II) vaccine, none of the associations was statistically associated with receipt of varicella vaccine. The diagnostic category of "rule out sepsis" still had a relative risk of 1.95 with P = 0.02. None of the children in the "rule out sepsis" category had positive bacteriologic cultures from any other normally sterile site. Because of the large number of gastroenteritis cases, we reviewed a random sample of 100 exposed and 100 unexposed cases. From this review no consistent time association or clustering of any of these events was seen in the exposed follow-up time interval. Only gastroenteritis and negative evaluations for sepsis were thought to be possibly associated with receipt of varicella vaccine. Although there was a statistically significant increased risk over the entire 30 day-period, there was no clustering of these events within the 30-day window.

CONCLUSION

In this study population of 89753 children and adults, the varicella vaccine (Oka strain, Merck) appeared to have a favorable safety profile. In addition rates of varicella-like rash and of breakthrough cases were both low and consistent with the rates observed in prelicensure studies.

National trends in the mortality of children with sickle cell disease, 1968 through 1992

OBJECTIVES

This paper describes national trends in mortality of children with sickle cell disease and the settings in which death occurred.

METHODS

United States death certificate data from 1968 through 1992 were used to calculate mortality rates of Black children with sickle cell disease 1 to 14 years old. Deaths from trauma, congenital anomalies, and perinatal conditions were excluded.

RESULTS

Between 1968 and 1992, mortality rates of Black children with sickle cell disease decreased 41% for 1- to 4-year-olds, 47% for 5- to 9-year-olds, and 53% for 10- to 14-year-olds. During 1986 through 1992, children who died before hospital admission accounted for 41% of deaths among 1- to 4-year-olds, 27% among 5- to 9-year-olds, and 12% among 10- to 14-year-olds.

CONCLUSIONS

Survival of Black children with sickle cell disease has improved markedly since 1968. A substantial proportion of deaths continue to occur prior to hospital admission. Trends in sickle cell mortality can be monitored inexpensively with death-certificate data.

The utilization of antidepressants--a key issue in the prevention of suicide: an analysis of 5281 suicides in Sweden during the period 1992-1994

Antidepressants detected by the National Department of Forensic Chemistry in 5281 suicides in Sweden during the period 1992-1994 were related to data on usage expressed in person-years of exposure. Antidepressants were detected in 874 subjects (16.5%). In relation to their use, fluvoxamine, citalopram, moclobemide, mianserin and trimipramine were found more often than the reference drug, amitriptyline (i.e. over-risks). Toxic concentrations of antidepressants were detected in 232 subjects (4.4%). Most people committing suicide were not taking antidepressants immediately before their death, even though 40-85% may have been depressed. Undertreatment and therapeutic failure are the main problems with antidepressants, not the risk of using antidepressants in overdose. Comparisons of new antidepressants should focus on efficacy in relation to reference tricyclics. The huge increase in the use of antidepressants in Sweden since 1990-1991 has been paralleled by a significant decrease in suicide rates.

Tuberculosis and HIV infection among female inmates in São Paulo, Brazil: a prospective cohort study

Prison populations are at increased risk of both human immunodeficiency virus (HIV) and Mycobacterium tuberculosis infections, but among female inmates information on such risks remains scarce, especially in developing countries. Between October 1992 and November 1993, 350 women incarcerated at a prison in São Paulo, Brazil, were prospectively evaluated for HIV and M. tuberculosis infection and disease. Among them, 87 (25%) were HIV seropositive, and 20 (5.7%) had tuberculosis (TB). During the incarceration period, the purified protein derivative test conversion rate was 29% for HIV-positive and 32% for HIV-negative women. However, the incidence of TB was 9.9 per 100 person-years for HIV-positive and 0.7 per 100 person-years of incarceration for HIV-negative women (p < 0.0001). A multivariate analysis indicated that HIV infection (p < 0.0001) and incarceration time < 12 months (p < 0.05) were each associated with TB. These findings indicate that new transmissions of M. tuberculosis infection are common among female inmates and that HIV-infected women are more likely to acquire active disease during the first 12 months of incarceration. Because of their role in childbearing and care female inmates are an important potential source of transmission of M. tuberculosis, and new strategies to control the spread of TB in prisons need to be developed.

The incidence of depression in new users of beta-blockers and selected antihypertensives

We studied the occurrence of depression in new users of propranolol (n = 704), other beta-blockers (n = 587), angiotensin-converting enzyme inhibitors (n = 976), calcium channel blockers (n = 742), and diuretics (n = 773) in the Harvard Community Health Plan population. The period of the study was from April 1988 to June 1991. All study subjects were followed for new or newly recurrent depression for up to 6 months after receiving their first study prescription. Case status was confirmed by blinded medical record review. We found 10 cases of depression that met DSM-III-R criteria ("major depression") and an additional 18 cases that had one or more symptoms consistent with depression ("minor depression"). Rates of major depression in users of beta-blockers and users of non-beta-blocker study drugs were 5.8 per 1000 person-years of exposure and 9.6 per 1000 person-years, respectively. None of the cases of major depression was propranolol associated. Rates of major or minor depression (combined) in users of beta-blockers and users of non-beta-blockers were 20.2 per 1000 person-years and 25.2 per 1000 person-years, respectively. The age- and sex-adjusted relative risk of major or minor depression associated with the use of beta-blockers compared to non-beta-blockers was 0.8 (95% CI, 0.3-1.9). The relative risk associated with propranolol compared to non-beta-blockers was also 0.8 (95% CI, 0.1-2.7). Therefore, depression occurred no more frequently in beta-blocker users than in other members of the study base.

Is there an excess in maternal transmission of NIDDM?

Family studies have demonstrated that there is a strong genetic component to the aetiology of non-insulin-dependent diabetes mellitus (NIDDM), although the mode of inheritance is unknown. A number of recent family history studies, including one in Mexican Americans, have suggested that there is an excess of maternal transmission of NIDDM. Family history studies are subject to various types of bias, however, and the potential for bias in many of these studies has not been thoroughly evaluated. We therefore tested the hypothesis that diabetes is more likely to be transmitted from mothers than from fathers using data collected from a large family study of low-income Mexican Americans in San Antonio, Texas. The parents and offspring from 318 different nuclear families attended our medical clinic, where they received a 2-h oral glucose test. Diabetes was diagnosed on the basis of World Health Organization criteria. The sibships were classified into diabetic sibships (at least one sibling in the sibship was diabetic; n = 54) and non-diabetic siblings (no diabetic siblings; n = 264). The prevalence of diabetes among mothers of diabetic siblings was 61.4% (27 of 44) compared to 64.3% (18 of 28) among fathers of diabetic siblings (rate ratio = 0.95; 95% confidence interval; 0.51-1.84). For the non-diabetic sibships, the prevalence of diabetes was 31.7% (78 of 246) and 28.9% (37 of 128) among mothers and fathers, respectively (rate ratio = 1.09; 95% confidence interval: 0.73-1.67). These data provide no evidence for an excess maternal transmission of diabetes in Mexican Americans.

Cancer incidence among firefighters in Seattle and Tacoma, Washington (United States)

In order to determine if exposure to carcinogens in fire smoke increases the risk of cancer, we examined the incidence of cancer in a cohort of 2,447 male firefighters in Seattle and Tacoma, (Washington, USA). The study population was followed for 16 years (1974-89) and the incidence of cancer, ascertained using a population-based tumor registry, was compared with local rates and with the incidence among 1,878 policemen from the same cities. The risk of cancer among firefighters was found to be similar to both the police and the general male population for most common sites. An elevated risk of prostate cancer was observed relative to the general population (standardized incidence ratio [SIR] = 1.4, 95 percent confidence interval [CI] = 1.1-1.7) but was less elevated compared with rates in policemen (incidence density ratio [IDR] = 1.1, CI = 0.7-1.8) and was not related to duration of exposure. The risk of colon cancer, although only slightly elevated relative to the general population (SIR = 1.1, CI = 0.7-1.6) and the police (IDR = 1.3, CI = 0.6-3.0), appeared to increase with duration of employment. Although the relationship between firefighting and colon cancer is consistent with some previous studies, it is based on small numbers and may be due to chance. While this study did not find strong evidence for an excess risk of cancer, the presence of carcinogens in the firefighting environment warrants periodic re-evaluation of cancer incidence in this population and the continued use of protective equipment.

Use of antidepressants among people committing suicide in Sweden

OBJECTIVE

To analyse the outcome of depression in the Swedish population as reflected by the detection of antidepressants in a national forensic toxicological screening programme of unnatural deaths.

DESIGN

Antidepressants detected by the National Laboratory of Forensic Chemistry were related to data on use expressed in person years of exposure.

SUBJECTS

All 7000 cases of unnatural death with results from forensic toxicological screening in 1990-1; this included 3400 (85%) of the 4000 cases of suicide in Sweden.

MAIN OUTCOME MEASURES

Number of findings of antidepressants in the screening programme and number of findings of different antidepressants in relation to their use.

RESULTS

Antidepressants were found in 585 screening tests, corresponding to 542 subjects or less than 16% of the 3400 cases of suicide. Newer, less toxic antidepressants were found more often than the older compounds. Toxic concentrations of antidepressants were found in only 190 cases (5.6%).

CONCLUSION

A consistent finding in surveys of suicide is that about half of the patients who commit suicide are depressed. The current data suggest that most depressed patients who commit suicide are not taking antidepressants immediately before death. Therapeutic failure may be a greater problem with antidepressants than toxicity as the results did not indicate any advantage of the newer, less toxic antidepressants. All causes of death should be included in risk analyses, thereby providing an estimate of effectiveness as well as toxicity of antidepressants.

Long-term malaria prophylaxis with weekly mefloquine

The spread of chloroquine-resistant Plasmodium falciparum malaria has led to increased use of mefloquine prophylaxis by US Peace Corps volunteers in sub-Saharan Africa. We compared long-term mefloquine with other drug regimens for effectiveness and tolerance. The incidence of Plasmodium falciparum infections and of adverse reactions was compared in Peace Corps volunteers who took chloroquine weekly, mefloquine weekly, mefloquine every other week, or weekly chloroquine plus daily proguanil. Weekly mefloquine was 94% more effective than chloroquine (95% CI 86% to 97%), 86% more effective than chloroquine plus proguanil (95% CI 67% to 94%), and 82% more effective than prophylaxis with mefloquine when taken every other week (95% CI 68% to 90%). No serious adverse reactions were observed. Mild adverse events were equally frequent in mefloquine users and chloroquine users, and the frequency of these events declined with increasing duration of prophylaxis. Mefloquine is an effective and well-tolerated drug for prophylaxis of malaria by short-term and long-term travellers.

A bibliography and comments on the use of statistical models in epidemiology in the 1980s

This paper reviews developments in statistical modelling in epidemiology in the 1980's, with emphasis on cohort and case-control studies. The central roles of the logistic and proportional hazard models are highlighted, and it is shown how these models lead to a deeper understanding of classical designs and methods of analysis as well as to efficient new designs and analytical procedures. The important area of model misspecification is discussed, including the problems of omitted latent structure, mis-modelling of available measurements, missing data and errors in measurements. Various designs motivated by the logistic model are illustrated numerically, and designs based on the proportional hazards model are discussed, as are papers on sample size determination. There are brief introductions to the literature on other topics, including attributable risk, disease clustering, family studies and genetics, analysis of disease incidence data, infectious disease, longitudinal data, screening and miscellaneous related topics in statistics. An extensive bibliography is indexed according to the outline of the paper.

Pharmacoepidemiology in pre-approval clinical trial safety monitoring

Applications of pharmacoepidemiology to the pre-approval safety monitoring of investigational drugs are examined and compared with the post-approval applications. Pre-approval epidemiologic assessments of drug safety can complement clinical assessments of rare, serious adverse events occurring in open label uncontrolled clinical trials. Applications include retrospective identification of patient risk factors for adverse drug events and incidence comparisons based on historical controls. While similar in concept to applications of pharmacoepidemiology in the analyses of uncontrolled open label post-marketing surveillance studies, answers are often needed in days rather than weeks and the emphasis is on whether human studies can continue. Advance planning of historical data sources is essential as is an effective clinical trial data management system. Implications for methodologic research and for drug development are discussed.

Acetazolamide-associated aplastic anaemia

Eleven cases of acetazolamide-associated aplastic anaemia were reported in Sweden during a 17-year period. There were six women and five men with a median age of 71 years (range 63-85 years). The median dose of acetazolamide was 500 mg, and the median duration of treatment was 3 months (range 2-71 months). Ten of the eleven patients died, all within 8 weeks after detection of their aplastic anaemia. The relative risk of developing aplastic anaemia when taking acetazolamide was 13.3 (95% confidence limits (CL); 6.8-25.3). The estimated incidence of reported acetazolamide-associated aplastic anaemia is approximately one in 18,000 patient years. The results strongly indicate that acetazolamide treatment is associated with a substantial increase in the risk of developing aplastic anaemia.

Invasive Haemophilus influenzae infections among children in Rochester, Minnesota: a population-based study

The incidence of invasive Haemophilus influenzae infections was studied among all residents of Rochester, Minnesota, under 20 years of age during the period 1975-1983. The incidence in children under 5 years of age was 115 and 64 cases per 100,000 children per year for all invasive cases and for meningitis, respectively. These rates are among the highest reported for an essentially all-Caucasian population in the United States, while the ratio of meningitis to non-meningitis cases (1.1:1) is among the lowest. This suggests possible under-ascertainment of non-meningitis cases in previous US studies of Haemophilus influenzae infections.

Trimethoprim-sulphamethoxazole-associated blood dyscrasias. Ten years' experience of the Swedish spontaneous reporting system

During the 10-year period 1976-1985, a total of 154 cases of blood dyscrasia were reported in Sweden which were evaluated as having a probable or possible causal relationship with trimethoprim-sulphamethoxazole (T-SM). There were 61 cases of leucopenia (of which 16 had agranulocytosis), 28 cases of thrombocytopenia, and two of non-haemolytic anaemia. There were also 32 cases of bicytopenia and 31 cases of tricytopenia. The median age varied from 38 years in the leucopenia group to 81 years in those with tricytopenia. The overall fatality rate was 17%, ranging from 2% in the group with mild leucopenia to 52% in the group with tricytopenia. In relation to sales and prescription data, the overall incidence of reported T-SM blood dyscrasias was 5.3 per million defined daily doses, and among out-patients the incidence was one case per 18,000 prescriptions. Thus the overall incidence of any blood reaction to T-SM appears to be low. In relation to prescription data, elderly people were overrepresented among the serious reactions.

Long-term sick leave and disability pensioning due to back disorders of tractor drivers exposed to whole-body vibration

In a historical 11-year follow-up study, disability pensioning and the incidence of the first sick leave of 4 weeks or longer due to back disorders has been investigated in a group of drivers exposed to whole-body vibration (WBV), mainly of agricultural tractors. The reference group comprised workers not or only slightly exposed to WBV from the same and another company. The vibration exposure of the tractor drivers was roughly around the ISO-2631 fatigue-decreased proficiency limit. The incidence of a first long-term sick leave due to back disorders was 3 per 100 person-years in tractor drivers and 2 per 100 person-years in the entire reference group. However, the incidence was not substantially increased in tractor drivers when the referents comprised only those working at the same company, suggesting selection bias. The highest relative risk (ca. 3) was found for long-term sick leave due to intervertebral disc disorders and this risk seemed to increase with the received WBV dose. That especially the incidence of intervertebral disc disorders increased, agrees well with findings in other studies. Data on disability pensioning due to back disorders are too scarce to be conclusive, but a trend towards younger disablement in tractor drivers is evident. Exposure to WBV, together with twisted posture and prolonged sitting, are considered to be responsible for the increased incidences observed in tractor drivers.

Fatal upper gastrointestinal hemorrhage or perforation among users and nonusers of nonsteroidal anti-inflammatory drugs in Saskatchewan, Canada 1983

We report a cohort study of fatal upper GI hemorrhage and/or perforation in relation to use of nonsteroidal anti-inflammatory drugs (NSAIDs) among the one million residents of Saskatchewan Canada in 1983. All hospitalized cases of GI hemorrhage and/or perforation with a fatal outcome were identified using the records linkage system of the Saskatchewan Department of Health. Discharge summaries and autopsy records were reviewed to select the cases of upper GI hemorrhage or upper GI perforation and to exclude cases in which known risk factors were present. The 134,060 residents who filled one or more prescriptions for an NSAID in 1983 were identified and individually linked to their hospital records by patient identification number. The age- and gender-specific incidence of fatal upper GI hemorrhage and/or perforation in the absence of risk factors in users was compared to that in nonusers, controlling for recent history of upper GI disease. Fatal upper GI hemorrhage or perforation in temporal association with NSAIDs is extremely rare in persons younger than 75 years of age. No temporally-related cases occurred in male NSAID users age 75 and older, but NSAID usage in this group was limited. Among women age 75 and older, the rate in users was higher than in nonusers, with the highest rate being in female NSAID users age 75 and older with a recent history of upper GI disease. Total mortality among women age 75 and older was slightly lower among users than among nonusers. Physicians who prescribe NSAIDs to patients age 75 and older should be aware of the potential risks, particularly in those with predisposing factors such as a history of upper GI disease.