Die Schweizerische HIV-Kohortenstudie – Ressource für Epidemiologie und Prävention

Die epidemiologische Situation in der Schweiz war gekennzeichnet durch einen deutlichen Rückgang in den Aidsfällen und HIV-assoziierten Todesfällen nach der Einführung der hochaktiven antiretroviralen Therapie (HAART). Die Neuinfektionen haben jedoch in den letzten Jahren wieder zugenommen und die Aidsfälle und Todesfälle nahmen nicht weiter ab. Wir diskutieren beispielhaft drei Fragenkomplexe, die für Epidemiologie und Public Health relevant und aktuell sind: (i) den Gebrauch von SHCS Daten in der Interpretation der Statistiken des Bundesamts für Gesundheit, die auf den Meldungen der Ärzte und Laboratorien beruhen; (ii) Analysen des Zugangs zu und Gebrauch von HAART in verschiedenen Patientengruppen, zum Beispiel Einwanderer, Patienten aus den unteren Sozialschichten und Patienten, die sich durch intravenösen Drogengebrauch mit HIV infiziert hatten; und (iii) die Erfassung von Trends im Sexualverhalten der Studienteilnehmer, inklusive Kondomgebrauch mit stabilen Partnern und bei Gelegenheitskontakten. Anhand dieser Beispiele versuchen wir zu zeigen, dass die SHCS eine wichtige Ressource für die epidemiologische Analyse und Überwachung der Epidemie und die Planung von Public Health Maßnahmen darstellt.

Long-Term Virological Response to Multiple Sequential Regimens of Highly Active Antiretroviral Therapy for HIV Infection

Objective

Information about the virological response to sequential highly active antiretroviral therapy (HAART) for HIV infection is limited. The virological response to four consecutive therapies was evaluated in the Swiss HIV Cohort.

Design

Retrospective analysis in an observational cohort.

Methods

1140 individuals receiving uninterrupted HAART for 4.8 ±0.6 years were included. The virological response was classified as success (<400 copies/ml), low-level (LF: 400–5000 copies/ml) or high-level failure (HF: >5000 copies/ml). Potential determinants of the virological response, including patient demographics, treatment history and virological response to previous HAART regimens were analysed using survival and logistic regression analyses.

Results

40.1% failed virologically on the first (22.0% LF; 18.1% HF), 35.1% on the second (14.2% LF; 20.9% HF), 34.2% on the third (9.9% LF; 24.3% HF) and 32.7% on the fourth HAART regimen (9% LF; 23.7% HF). Nucleoside pre-treatment (OR: 2.34; 95% CI: 1.67–3.29) and low baseline CD4 T-cell count (OR: 0.79/100 cells rise; 95% CI: 0.72–0.88) increased the risk of HF on the first HAART. Virological failure on HAART with HIV-1 RNA levels exceeding 1000 copies/ml predicted a poor virological response to subsequent HAART regimens. A switch from a protease inhibitor- to a non-nucleoside reverse transcriptase inhibitor-containing regimen significantly reduced the risk of HF. Multiple switches of HAART did not affect the recovery of CD4 T lymphocytes.

Conclusion

Multiple sequential HAART regimens do not per se reduce the likelihood of long-term virological suppression and immunological recovery. However, early virological failure increases significantly the risk of subsequent unfavourable virological responses. The choice of a potent initial antiretroviral drug regimen is therefore critical. This study has been presented in part at the 10th Conference on Retroviruses & Opportunistic Infections. Boston, Mass., USA, 2003. Abstract #571.

Effect of Individual Cognitive Behaviour Intervention on Adherence to Antiretroviral Therapy: Prospective Randomized Trial

Objective

A high level of adherence to antiretroviral therapy is required for complete suppression of HIV replication, immunological and clinical effectiveness. We investigated whether cognitive behaviour therapy can improve medication adherence.

Design

Prospective randomized 1-year trial.

Setting

Collaboration of HIV university outpatient clinic and psychotherapists in private practice.

Participants

60 HIV-infected persons on stable anti-retroviral combination therapy and viral load below 50 copies/ml.

Intervention

Cognitive behaviour intervention in individual patients, in addition to standard of care.

Main outcome measures

Feasibility and acceptance of intervention; adherence to therapy assessed using medication event monitoring system (MEMS) and self-report questionnaire; virological failure; psychosocial measures.

Results

The median number of sessions for cognitive behaviour intervention per patient during the 1-year trial was 11 (range 2–25). At months 10–12, mean adherence to therapy as assessed using MEMS was 92.8% in the intervention and 88.9% in the control group ( P=0.2); the proportion of patients with adherence ≥95% was 70 and 50.0% ( P=0.014), respectively. While there was no significant deterioration of adherence during the study in the intervention arm, adherence decreased by 8.7% per year ( P=0.006) in the control arm. No differences between the intervention group and standard of care group were found regarding virological outcome. Compared with the control group, participants in the intervention group perceived a significant improvement of their mental health during the study period.

Conclusions

Cognitive behavioural support in addition to standard of care of HIV-infected persons is feasible in routine practice, and can improve medication adherence and mental health.

Impact of Genotypic Resistance Testing on Selection of Salvage Regimen in Clinical Practice

Objective

To determine whether genotypic resistance testing leads to selection of more potent drug regimens when compared to regimens based on treatment history only.

Design

Prospective, tertiary care centre-based study. Patients: One-hundred-and-forty-five HIV-infected adults on stable antiretroviral therapy (ART) for >6 months experiencing virological failure.

Methods

The physicians’ decision-making process when choosing a salvage regimen was prospectively documented: at time of virological failure, on ‘failing ART’, genotyping was performed and a hypothetical ‘clinical expert ART’ based upon patient's drug history was documented. Subsequently, data on resistance mutations, rating by a decision support software and drug history were used to define ‘genotyping ART’. After discussion with the patient, final treatment, ‘new personalized ART’ was chosen and prescribed. To compare the relative potency of the four ART regimens in a standardized manner, a resistance score ranging from 1 (best) to 8 (worst) based on drug ranking by decision support software was attributed to each ART regimen. Virological and immunological outcomes were analysed based on the magnitude of the resistance score.

Results

Median follow-up was 1.5 years. In all 145 patients, median resistance scores for the stepwise selected ART regimens were: ‘failing ART’: 4.5, ‘clinical expert ART’: 1.8, ‘genotyping ART’: 1.5 and ‘new personalized ART’: 2. The latter was 1.5 in patients who effectively switched to ‘new personalized ART’ ( n=89). Lower resistance scores translated into significantly improved virological response after initiation of ‘new personalized ART’. In multivariable analysis, lower resistance scores, lower baseline HIV RNA levels and use of novel antiretroviral drugs were associated with the probability of reducing plasma viraemia to <50 copies/ml. Conclusions: This study suggests that treatment choices including genotype and decision support software were virologically superior to those based on drug history only.

Male circumcision for prevention of heterosexual acquisition of HIV in men

BACKGROUND

The findings from observational studies, reviews and meta-analyses, supported by biological theories, that circumcised men appear less likely to acquire human immunodeficiency virus (HIV) has contributed to the recent ground swell of support for considering male circumcision as a strategy for preventing sexually acquired infection. We sought to elucidate and appraise the global evidence from published and unpublished studies that circumcision can be used as an intervention to prevent HIV infection.

OBJECTIVES

1) To assess the evidence of an interventional effect of male circumcision for preventing acquisition of HIV-1 and HIV-2 by men through heterosexual intercourse 2) To examine the feasibility and value of performing individual person data (IPD) meta-analysis

SEARCH STRATEGY

We searched online for published and unpublished studies in The Cochrane Library (issue 2, 2002), MEDLINE (April 2002), EMBASE (February 2002) and AIDSLINE (August 2001). We also searched databases listing conference abstracts, scanned reference lists of articles and contacted authors of included studies.

SELECTION CRITERIA

We searched for randomized and quasi-randomized controlled trials of male circumcision or, in their absence, observational studies that compare acquisition rates of HIV-1 and HIV-2 infection in circumcised and uncircumcised heterosexual men.

DATA COLLECTION AND ANALYSIS

Independent reviewers selected studies, assessed study quality and extracted data. We stratified studies based on study design and on whether they included participants from the general population or high-risk groups (such as patients treated for sexually transmitted infections). We expressed findings as crude and adjusted odds ratios (OR) together with their 95% confidence intervals (CI) and conducted a sensitivity analysis to explore the effect of adjustment on study results. We investigated whether the method of circumcision ascertainment influenced study outcomes.

MAIN RESULTS

We identified no completed randomized controlled trials. Three randomized controlled trials are currently underway or commencing shortly. We found 34 observational studies: 16 conducted in the general population and 18 in high-risk populations. It seems unlikely that potential confounding factors were completely accounted for in any of the included studies. In particular, important risk factors, such as religion and sexual practices, were not adequately accounted for in many of the included studies. General population study results:The single cohort study (N = 5516) showed a significant difference in HIV transmission rates between circumcised and uncircumcised men [OR = 0.58; 95% CI: 0.36 to 0.96]. Results for the 14 cross-sectional studies were inconsistent, with point estimates for unadjusted odds ratios varying between 0.28 and 1.73. Six studies had statistically significant results, four in the direction of benefit and two in the direction of harm. The test for heterogeneity between the cross-sectional studies was highly significant (chi-square = 77.59; df = 13; P-value < 0.00001). Nine studies reported adjusted odds ratios with eight in the direction of benefit, ranging from 0.26 to 0.80. Use of adjusted results tended to show stronger evidence of an association although they remained heterogenous (chi-square = 75.2; df = 13; P-value < 0.00001). Only one case-control study was found (N = 51) which had a non-significant result [OR = 1.90; 95% CI: 0.50 to 7.20]. High-risk group study results:The four cohort studies identified found a protective effect from circumcision with point estimates for unadjusted odds ratios varying from 0.10 to 0.39. Two of these studies had statistically significant results. Two studies reported adjusted odds ratios, both protective with one being significant. The chi-square test for between-study heterogeneity was not significant (chi-square = 5.21; df = 3; P-value = 0.16). All eleven cross-sectional studies reporting unadjusted results found benefit from circumcision, eight of which had statistically significant results. Estimates of effnal studies reporting unadjusted results found benefit from circumcision, eight of which had statistically significant results. Estimates of effect varied from an unadjusted odds ratio of 0.10 to 0.66. Between-study heterogeneity was significant with the chi-square = 29.77; df = 10; P-value = 0.0009. Four of these studies reported adjusted odds ratios ranging from 0.20 to 0.59 and all were significant. One additional cross-sectional study only reported an adjusted odds ratio in the direction of benefit which was statistically significant. All three case-control studies found a protective effect of circumcision on HIV status, two being statistically significant. Point estimates varied from unadjusted odds ratios of 0.37 to 0.88. One reported an adjusted odds ratio showing a significant protective effect. Adverse effects:No studies reported on the adverse effects of circumcision. In most studies, circumcision had taken place during childhood or adolescence before the studies commenced.

REVIEWER'S CONCLUSIONS

We found insufficient evidence to support an interventional effect of male circumcision on HIV acquisition in heterosexual men. The results from existing observational studies show a strong epidemiological association between male circumcision and prevention of HIV, especially among high-risk groups. However, observational studies are inherently limited by confounding which is unlikely to be fully adjusted for. In the light of forthcoming results from RCTs, the value of IPD analysis of the included studies is doubtful. The results of these trials will need to be carefully considered before circumcision is implemented as a public health intervention for prevention of sexually transmitted HIV.

Do all men attending departments of genitourinary medicine need to be screened for non-gonococcal urethritis?

We investigated the influence of symptoms and signs on the detection of Chlamydia trachomatis, Mycoplasma genitalium and Ureaplasma urealyticum organisms (ureaplasmas) in men with non-gonococcal urethritis (NGU). Two hundred and forty-two men attending the Jefferiss Wing at St Mary's Hospital for a sexual health assessment were evaluated, of whom 169 had NGU. Urethral inflammation was diagnosed if there were either > or =5 polymorphonuclear leucocytes (PMNLs) per high-power field (HPF) in five or more microscope fields of a Gram-stained urethral smear, or > or =10 PMNLs per HPF in five or more fields of a Gram-stained thread from 15-20 mL of a first-passed urine (FPU) specimen. C. trachomatis was diagnosed by direct immunofluoresence, M. genitalium by a polymerase chain reaction assay and ureaplasmas by culture. On multivariate analysis, to control for potential confounding by age, ethnicity, sexual lifestyle and co-infection, an urethral discharge remained significantly associated with the detection of C. trachomatis and M. genitalium in men with acute urethritis [OR 12.3, 95% CI (2.39-63.5) and OR 35.2, 95% CI (3.9-319.6), respectively], but dysuria or penile irritation did not. The detection of ureaplasmas was not associated with any clinical feature. In addition, on multivariate analysis men with NGU who were either symptomatic or had an observable discharge were more likely to have C. trachomatis or M. genitalium detected [(OR 6.92, 95% CI 1.41-33.9) and (OR 5.18, 95% CI 0.99-27.1), respectively], but not ureaplasmas (OR 1.19, 95% CI 0.33-4.35). The findings suggest that in men with acute NGU, symptoms or signs, and in particular a urethral discharge, are associated with the detection of C. trachomatis and M. genitalium, but not ureaplasmas. Currently, there is no precise answer to the question of whether all men attending a GUM clinic need to be screened for NGU, but if clinically asymptomatic NGU is found not to be associated with a sexually transmitted pathogen, the UK clinical guidelines requiring the preparation of a urethral smear from such men would need to be revised.

Subgroup analyses in randomised controlled trials: quantifying the risks of false-positives and false-negatives

BACKGROUND

Subgroup analyses are common in randomised controlled trials (RCTs). There are many easily accessible guidelines on the selection and analysis of subgroups but the key messages do not seem to be universally accepted and inappropriate analyses continue to appear in the literature. This has potentially serious implications because erroneous identification of differential subgroup effects may lead to inappropriate provision or withholding of treatment.

OBJECTIVES

(1) To quantify the extent to which subgroup analyses may be misleading. (2) To compare the relative merits and weaknesses of the two most common approaches to subgroup analysis: separate (subgroup-specific) analyses of treatment effect and formal statistical tests of interaction. (3) To establish what factors affect the performance of the two approaches. (4) To provide estimates of the increase in sample size required to detect differential subgroup effects. (5) To provide recommendations on the analysis and interpretation of subgroup analyses.

METHODS

The performances of subgroup-specific and formal interaction tests were assessed by simulating data with no differential subgroup effects and determining the extent to which the two approaches (incorrectly) identified such an effect, and simulating data with a differential subgroup effect and determining the extent to which the two approaches were able to (correctly) identify it. Initially, data were simulated to represent the 'simplest case' of two equal-sized treatment groups and two equal-sized subgroups. Data were first simulated with no differential subgroup effect and then with a range of types and magnitudes of subgroup effect with the sample size determined by the nominal power (50-95%) for the overall treatment effect. Additional simulations were conducted to explore the individual impact of the sample size, the magnitude of the overall treatment effect, the size and number of treatment groups and subgroups and, in the case of continuous data, the variability of the data. The simulated data covered the types of outcomes most commonly used in RCTs, namely continuous (Gaussian) variables, binary outcomes and survival times. All analyses were carried out using appropriate regression models, and subgroup effects were identified on the basis of statistical significance at the 5% level.

RESULTS

While there was some variation for smaller sample sizes, the results for the three types of outcome were very similar for simulations with a total sample size of greater than or equal to 200. With simulated simplest case data with no differential subgroup effects, the formal tests of interaction were significant in 5% of cases as expected, while subgroup-specific tests were less reliable and identified effects in 7-66% of cases depending on whether there was an overall treatment effect. The most common type of subgroup effect identified in this way was where the treatment effect was seen to be significant in one subgroup only. When a simulated differential subgroup effect was included, the results were dependent on the nominal power of the simulated data and the type and magnitude of the subgroup effect. However, the performance of the formal interaction test was generally superior to that of the subgroup-specific analyses, with more differential effects correctly identified. In addition, the subgroup-specific analyses often suggested the wrong type of differential effect. The ability of formal interaction tests to (correctly) identify subgroup effects improved as the size of the interaction increased relative to the overall treatment effect. When the size of the interaction was twice the overall effect or greater, the interaction tests had at least the same power as the overall treatment effect. However, power was considerably reduced for smaller interactions, which are much more likely in practice. The inflation factor required to increase the sample size to enable detection of the interaction with the same power as the overall effect varied with the size of the interaction. For an interaction of the same magnitude as the overall effect, the inflation factor was 4, and this increased dramatically to of greater than or equal to 100 for more subtle interactions of < 20% of the overall effect. Formal interaction tests were generally robust to alterations in the number and size of the treatment and subgroups and, for continuous data, the variance in the treatment groups, with the only exception being a change in the variance in one of the subgroups. In contrast, the performance of the subgroup-specific tests was affected by almost all of these factors with only a change in the number of treatment groups having no impact at all.

CONCLUSIONS

While it is generally recognised that subgroup analyses can produce spurious results, the extent of the problem is almost certainly under-estimated. This is particularly true when subgroup-specific analyses are used. In addition, the increase in sample size required to identify differential subgroup effects may be substantial and the commonly used 'rule of four' may not always be sufficient, especially when interactions are relatively subtle, as is often the case. CONCLUSIONS--RECOMMENDATIONS FOR SUBGROUP ANALYSES AND THEIR INTERPRETATION: (1) Subgroup analyses should, as far as possible, be restricted to those proposed before data collection. Any subgroups chosen after this time should be clearly identified. (2) Trials should ideally be powered with subgroup analyses in mind. However, for modest interactions, this may not be feasible. (3) Subgroup-specific analyses are particularly unreliable and are affected by many factors. Subgroup analyses should always be based on formal tests of interaction although even these should be interpreted with caution. (4) The results from any subgroup analyses should not be over-interpreted. Unless there is strong supporting evidence, they are best viewed as a hypothesis-generation exercise. In particular, one should be wary of evidence suggesting that treatment is effective in one subgroup only. (5) Any apparent lack of differential effect should be regarded with caution unless the study was specifically powered with interactions in mind. CONCLUSIONS--RECOMMENDATIONS FOR RESEARCH: (1) The implications of considering confidence intervals rather than p-values could be considered. (2) The same approach as in this study could be applied to contexts other than RCTs, such as observational studies and meta-analyses. (3) The scenarios used in this study could be examined more comprehensively using other statistical methods, incorporating clustering effects, considering other types of outcome variable and using other approaches, such as Bootstrapping or Bayesian methods.

Hyperlactatemia and antiretroviral therapy: the Swiss HIV Cohort Study

The prevalence, clinical presentation, and risk factors for hyperlactatemia among patients receiving antiretroviral therapy was determined during a 1-month period for patients in the Swiss HIV Cohort Study. Overall, 73 (8.3%) of 880 patients presented an increase in serum lactate of >1.1 times the upper normal limit (UNL). For 9 patients (1%), lactate elevation was moderate or severe (>2.2 times the UNL). Patients who presented with hyperlactatemia were more likely to be receiving stavudine with or without didanosine (odds ratio, 2.7; 95% confidence interval, 1.5-4.8), as compared with patients who received zidovudine-based regimens. The risk increased with increasing time receiving stavudine with or without didanosine. The association between hyperlactatemia and stavudine with or without didanosine was not biased by these medications being more recently available and, therefore, being given preferentially to patients who had prolonged use of nucleoside analog reverse-transcriptase inhibitors. Hyperlactatemia was associated with lipoatrophy, hyperlipidemia, and hyperglycemia. Age, sex, or stage of infection with human immunodeficiency virus were not predictive of hyperlactatemia. Determination of lactate levels may prove useful in the screening for mitochondrial toxicity.

Uses and abuses of meta-analysis

Meta-analysis, the statistical combination of results from several studies to produce a single estimate of the effect of a treatment, continues to attract controversy. We illustrate the potentials and pitfalls of meta-analysis of controlled clinical trials. Cumulative meta-analysis demonstrates that this technique could prevent delays in the introduction of effective treatments. Meta-analyses are, however, liable to numerous biases both at the level of the individual trial ('garbage in, garbage out') and the dissemination of trial results (publication bias). We argue that meta-analysis should be performed only within the framework of systematic reviews--that is, reviews prepared using a systematic approach to minimise bias and address the combinability of studies.

Early enteral feeding versus "nil by mouth" after gastrointestinal surgery: systematic review and meta-analysis of controlled trials

OBJECTIVE

To determine whether a period of starvation (nil by mouth) after gastrointestinal surgery is beneficial in terms of specific outcomes.

DESIGN

Systematic review and meta-analysis of randomised controlled trials comparing any type of enteral feeding started within 24 hours after surgery with nil by mouth management in elective gastrointestinal surgery. Three electronic databases (PubMed, Embase, and the Cochrane controlled trials register) were searched, reference lists checked, and letters requesting details of unpublished trials and data sent to pharmaceutical companies and authors of previous trials.

MAIN OUTCOME MEASURES

Anastomotic dehiscence, infection of any type, wound infection, pneumonia, intra-abdominal abscess, length of hospital stay, and mortality.

RESULTS

Eleven studies with 837 patients met the inclusion criteria. In six studies patients in the intervention group were fed directly into the small bowel and in five studies patients were fed orally. Early feeding reduced the risk of any type of infection (relative risk 0.72, 95% confidence interval 0.54 to 0.98, P=0.036) and the mean length of stay in hospital (number of days reduced by 0.84, 0.36 to 1.33, P=0.001). Risk reductions were also seen for anastomotic dehiscence (0.53, 0.26 to 1.08, P=0.080), wound infection, pneumonia, intra-abdominal abscess, and mortality, but these failed to reach significance (P>0.10). The risk of vomiting was increased among patients fed early (1.27, 1.01 to 1.61, P=0.046).

CONCLUSIONS

There seems to be no clear advantage to keeping patients nil by mouth after elective gastrointestinal resection. Early feeding may be of benefit. An adequately powered trial is required to confirm or refute the benefits seen in small trials.

Funnel plots for detecting bias in meta-analysis: guidelines on choice of axis

Asymmetry in funnel plots may indicate publication bias in meta-analysis, but the shape of the plot in the absence of bias depends on the choice of axes. We evaluated standard error, precision (inverse of standard error), variance, inverse of variance, sample size and log sample size (vertical axis) and log odds ratio, log risk ratio and risk difference (horizontal axis). Standard error is likely to be the best choice for the vertical axis: the expected shape in the absence of bias corresponds to a symmetrical funnel, straight lines to indicate 95% confidence intervals can be included and the plot emphasises smaller studies which are more prone to bias. Precision or inverse of variance is useful when comparing meta-analyses of small trials with subsequent large trials. The use of sample size or log sample size is problematic because the expected shape of the plot in the absence of bias is unpredictable. We found similar evidence for asymmetry and between trial variation in a sample of 78 published meta-analyses whether odds ratios or risk ratios were used on the horizontal axis. Different conclusions were reached for risk differences and this was related to increased between-trial variation. We conclude that funnel plots of meta-analyses should generally use standard error as the measure of study size and ratio measures of treatment effect.

Increased risk of wasting syndrome in HIV-infected travellers: prospective multicentre study

HIV-infected patients from moderate regions who travel in tropical countries may experience clinical disease progression due to exposure to bacteria (including mycobacteria), fungi and parasites. In the Swiss HIV Cohort Study we examined the hypothesis that travelling increases the risk of tuberculosis, wasting syndrome, cryptosporidiosis, isosporiasis, cryptococcosis, coccidiomycosis, histoplasmosis and Salmonella septicaemia. A total of 4549 participants were included (in 1988-98) of whom 596 (13.1%) travelled at least once. During 16,800 person-years of follow-up 231 patients developed at least 1 of the diseases of interest. Wasting syndrome was the only diagnosis significantly associated with travelling (hazard ratio 2.16, 95% confidence interval 1.09 to 4.30). The risk of wasting syndrome ('slim disease') should be taken into account when counselling HIV-infected patients intending to travel in tropical regions.