 |
| Figure 2 From HCTZ Article |
Dr. Kaur: Antihypertensive Efficacy of Hydrochlorothiazide
Click here to listen to a recording of Dr. Kaur's presentation.
Dr. Kaur presented a meta-analysis of studies comparing the efficacy of hydrochlorothiazide to that of several other antihypertensive classes as assessed by ambulatory blood pressure monitoring. Messerli et al. write that they undertook this study because, while HCTZ remains the most widely prescribed antihypertensive in the US and is recommended as a first-line therapy by the JNC-7, "little evidence is available regarding the efficacy and safety of HCTZ for the treatment of essential hypertension."
The meta-analysis was conducted according to established protocols (as outlined in the QUORUM guidelines) and analyzed using standard statistical techniques. The authors found that HCTZ displayed consistently lower efficacy in reducing blood pressure than all other classes against which it was tested (see Fig 2, above). They concluded, rather forcefully, that "because of such paltry antihypertensive efficacy and the lack of outcome data at these doses, physicians should refrain from prescribing HCTZ as initial antihypertensive therapy.
The point that outcome data on HCTZ in hypertension is sparse may come as a surprise to most internal medicine residents, who could be forgiven for assuming that such as prestigious practice guideline as the JNC-7 would be based on considerable evidence. However, as Dr. Flattery pointed out, the JNC-7 recommendation of "thiazide-type diuretics" is primarily based on the ALLHAT trial, a large study which documented the efficacy and long-term benefits of chlorthalidone, a different drug in the thiazide class.
Dr Wofsy and Dr. Berry expressed some dismay that the authors didn't appear to have taken industry funding into account when preparing their analysis. Although they did assess for publication bias, the simple facts are that most head-to-head drug trials are funded by the pharmaceutical industry, and that industry-funded drug studies are more likely to show positive results than studies with no funding by the manufacturer. Since HCTZ has been generic for years, one would expect the trials analyzed to have been funded by the manufacturers of the other drug, and it is therefore unsurprising (if you have any cynicism at all about free-market medicine) that it should be consistently outperformed.
The general consensus was that it would be premature to change practice based on this single meta-analysis. However, it will be worth remaining alert to this issue in the event that further trial data does become available. The recent ACCOMPLISH trial tested HCTZ against amlodipine, (although both were combined with benzepril) and demonstrated an increased incidence of cardiovascular morbidity in the HCTZ group; it may be that future research will move us away from HCTZ on more definitive grounds, if only towards chlorthalidone.
Dr. Patel: Changes in Diet and Lifestyle and Long-Term Weight Gain in Women and Men
Click here to listen to a recording of Dr. Patel's presentation.
Dr. Patel reviewed a meta-analysis of three large prospective observational cohort studies which tried to correlate specific foods and health behaviors with incremental gains in weight over time. Their results are summarized in Table 1:
The results did not exactly reveal a plethora of astonishing secrets; it turns out (mirabile dictu) that eating french fries and watching TV were, in fact, associated with greater weight gain than eating brown rice and starting to exercise. However, a few things do leap out at one, among them the seemingly disproportionate effects of french fries and yogurt in opposite directions. Regarding the former, Dr. Berry drew our attention to one of the comments on the article on the NEJM website, which hypothesizes that the correlation between fried potatoes and massive weight gain may be due to the fact that some of the fatty acids present in potato chips and french fries are precursors for chemicals active in the endogenous cannabinoid system; put simply, they may give one "the munchies". In considering the apparently protective effect of yogurt, the authors refer to recent work suggesting that obesity is correlated with in humans and can be induced in animal models by alteration of the colonic microbiome.
As Dr. Wofsy pointed out, one can only take prospective observational studies whose measurements are collected by self-report so seriously, since the potential for recall bias and the range of possible confounders are both immense. Dr. Hicks was concerned by the fidelity with which the study replicates the dietary mores of the contemporary educated middle-class. She also drew attention to numerous possible confounders, including the correlation of many of the behaviors observed to be associated with weight gain with depression, which was not measured in the study populations, and the fact that computers supplanted televisions as the main form of sedentary occupation for most people during the study period and that this represents a significant and unmeasured variable. She concluded by stressing that in primary care clinic, the important thing is to establish what healthy choices a particular patient can make within the range of social and economic possibilities available to her; not to try to bring her in line with the abstracted appetites and shopping
preferences of people who may not resemble the patient herself.
Dr. Sombredero: Effect of Culturally Tailored Diabetes Education in Ethnic Minorities With Type 2 Diabetes
Click here to listen to a recording of Dr. Sombredero's presentation.
Dr. Sombredero reviewed two studies, one a trial of a specific culturally-tailored diabetes education intervention, the other a meta-analysis of culturally tailored diabetes interventions in general. The meta-analysis did manage to demonstrate a positive pooled effect for culturally sensitive educational interventions.
However, it's not exactly clear why this would be best demonstrated by meta-analysis, since the things they're comparing (interventions targeted at populations ranging from Hispanics living in America to rural African American women to Portuguese-Canadians and British Pakistanis) are clearly not the same except insofar as they somehow involve "culture." Their approach seems loosely analogous to attempting to pool the effect sizes of surgery, medical therapy, and radio-active iodine therapy on multinodular hyperthyroidism. You would get a positive result, since they all work, but attempting to calculate an average effect would be meaningless since, in the end, any given patient gets only one. Similarly, a patient is either British Pakistani or rural African American; there's no conceptual way to "average" the influences of an intervention directed specifically at one or the other with reference to both.
That said, it seems pretty self-evident that culturally-tailored interventions are a good idea (that is, after all, what we do when we speak English to English-speaking patients, and it seems obvious that that works better than any alternative), and insofar as this study furthers their development it may be valuable as a political intervention, if not a scientific one.
No comments:
Post a Comment