Does Chocolate Make Acne Worse?

Smith RN, et al. J Am Acad Dermatol. 2007;57:247-256.

Currently prevailing scientific opinion denies any relationship between diet and acne: teenagers throughout the land rejoice over their freedom to eat as much pizza, popcorn, chocolate, and soda (the four food groups from age 14-20) as they wish. But the last word may not yet be in.

Acne is related to both androgen and insulin in a complex fashion. Probably the most commonly recognized syndrome of hyperinsulinemia and hyperandrogenism is PCOS (Polycystic Ovary Syndrome). The constellation of acne, oligomenorrhea, obesity, infertility, and hirsutism in women should lead to a search for PCOS. Since hyperinsulinemia and acne in PCOS are comorbid, the link between diet—which can lead to hyperinsulinemia—and acne becomes plausible.

Young adult subjects with acne (n = 54) were randomized to either a low-glycemic index diet (LGID) or a non-modified high carbohydrate diet (control) for 12 weeks. All acne medications were discontinued, but all participants in both groups used daily Cetaphil skin cleanser.

At the end of 12 weeks, acne lesion counts were significantly lower in the LGID compared to control. Similarly, insulin sensitivity had improved in the LGID group, but worsened in the control group. Finally, sex hormone binding globulin (SHBG) levels, which are responsible for how much free testosterone circulates, were reduced in the control group (producing more free testosterone, and greater proclivity to acne). Adrenal androgens were also lower in the LGID group. In response to the authors' closing statement "To our knowledge, this is the first study to demonstrate a therapeutic effect of dietary intervention on acne. "

Psoriasis Goes More Than Skin Deep

Kremers HM, et al. J Am Acad Dermatol. 2007;57:347-354.

It is increasingly recognized that some immune-mediated disorders have consequences much more far reaching than their primary tissue compartment pathology. For instance, we know that patients with rheumatoid arthritis have an increased risk of CV disease, pulmonary disease, and malignancy. A critical link between immune-mediated disease and atherosclerosis appears to be inflammation. Psoriasis has already signaled its ability to induce distant inflammation by its association with comorbid psoriatic arthritis.

Several population-based studies have found an increased risk of cardiovascular disease in persons with psoriasis. The UK's General Practice Research Database (n = 130,976) found a 3-fold increase in risk of MI for young adults with severe psoriasis. A Swedish study of hospitalized patients showed a greater mortality amongst psoriasis patients by 50% compared with expected.

There are several putative mechanisms for increased CV risk. Persons with psoriasis more commonly smoke, and are more commonly overweight and sedentary. Also, some of the therapeutic agents for psoriasis may increase CV risk. Finally, inflammatory markers associated with psoriasis may participate in the pathways of inflammation that lead to atherosclerosis. In the same way we have come to recognize that the presence of rheumatoid arthritis is worthy of being considered a "CV risk factor," clinicians may appreciate a newly identified association of psoriasis and CV disease.

Riding the Chromium Pendulum

Balk EM, et al. Diabetes Care. 2007;30:2154-2163.

Chromium at one time carried the moniker "Glucose Tolerance Factor" due to its critical role in glucose metabolism. In severe chromium deficiency states, diabetes not uncommonly ensues; severe chromium deficiency is, however, uncommon in the United States. Despite the uncertain role of chromium in persons of normal nutriture, sales of chromium indicate popularity amongst the general public: 6% of mineral supplement sales are attributable to chromium.

In an effort to provide clarification of the role of chromium supplements, Balk et al performed a systematic review of randomized controlled trials (n = 41 trials that met eligibility criteria). The discussion that follows pertains to diabetic or pre-diabetic subjects.

Although 11 of 14 studies found no effect or a statistically insignificant effect, the overall systematic review showed that chromium supplementation reduced A1c about 0.6%; there was also a modest favorable effect on fasting glucose: a reduction of about 18 mg%. There was some variability in response to chromium depending upon formulation. Chromium provided as brewers yeast had the greatest effect, followed by chromium picolinate, and lastly chromium chloride.

This systematic review finds favorable effects of chromium supplements. Because the studies which showed the most favorable impact were often of the least quality, the authors opine that this conclusion should not be viewed as definitive.

Size Matters … Sometimes

Gotzsche PC, Hrobjartsson, et al. JAMA. 2007;298(4):430.

Meta-analysis (META) is intended to provide greater strength of evidence by combining like populations of similar clinical trials in similar circumstances. The operative word here, of course, is "similar." Unfortunately, no two populations or sets of circumstances are identical, resulting in conclusions being drawn from large data sets which may not be reliable. Indeed, there are numerous examples of META based on huge population size that were ultimately disproved on the basis of a single large randomized controlled trial.

The reader would likely assume that when reviewing literature in reputable journals, reporting of the evidence would be accurate. According to Gotzsche, et al, readers should "approach such meta-analyses with caution."

Because not all trials use the same scale for measuring outcomes, standardized mean difference (SMD) is a tool utilized to make outcomes measured by a variety of metrics fit the same, uniform, comparative scale. Simple in theory and application, Gotzsche evaluated META to see whether SMD had been accurately applied.

When "rechecking" the analyses, the authors could not replicate results in 37% of meta-analyses they examined. Of 10 meta-analyses employing SMD, 70% were erroneous. The degrees of error in some meta-analyses were sufficient to reverse the conclusions of the investigation.

The authors call for greater vigilance by readers and reviewers in reference to meta-analyses which utilize SMD methodology.

Peripheral Arterial Disease and Antithrombotic Therapy: Sometimes Less is More

The Warfarin Antiplatelet Vascular Evaluation Trial Investigators. N Engl J Med. 2007;357:217-227.

It is increasingly recognized that vasculopathy is usually not compartmentalized. Persons with CAD are at greater risk of stroke; those with peripheral arterial disease (PAD) are at increased risk of stroke and MI, and so on.

Aspirin reduces risk of CVD endpoints in persons with PAD, and at least one trial (CAPRIE) has shown that clopidogrel is modestly more effective than aspirin (absolute risk reduction 0.5%) in this regard. Recent years have seen trials in atrial fibrillation, post-stroke, stable coronary disease, and other mixed populations to determine whether combination therapies (eg, aspirin plus clopidogrel, warfarin plus antiplatelet) might be more beneficial than monotherapy. With the exception of acute coronary syndromes, combination therapies have generally been no more effective than monotherapy (for secondary prevention of stroke, the combination proprietary product aspirin-extended release dipyridamole is an exception to this rule).

The Warfarin Antiplatelet Vascular Evaluation Trial Investigators studied PAD patients (n = 2161) randomly assigned to either antiplatelet therapy alone (ASA, clopidogrel, or ticlopidine) vs an antiplatelet agent plus warfarin, targeted to an INR of 2-3, for a mean of 35 months. The two coprimary outcomes were MI, stroke, or CV death (composite) and MI, stroke, CV death, or need for surgical vascular intervention (composite).

Combination therapy did NOT show any statistically significant reduction in either coprimary endpoint. Additionally, however, combination therapy was associated with a three-fold greater incidence of life-threatening bleeding (4% vs 1.2%). As has been encountered in other vascular disease states, sometimes less is more.

Osteoporosis in Men: Cost-effectiveness Issues

Schousboe JT, et al. JAMA. 2007;298(6):629-637.

Many clinicians probably immediately think "woman" when they hear the word "osteoporosis," despite the fact that 1/3 of all hip fractures happen in men. Similarly, literature to support the use of interventions to prevent osteoporotic fracture in men has lagged behind the data set in women. The current report by Schousboe, et al will help add gender balance to the knowledge base.

Unfortunately, there has been only one prior cost-effectiveness modeling study of osteoporosis (OSPS) in men, and this study did not include the cost of bone densitometry (DEXA) in their model. The current study assumed that once OSPS was detected, bisphosphonate therapy would be initiated at an average cost of $1000 per year. DEXA cost was set at $82 (current Medicare reimbursement rate). The currently accepted threshold for an intervention to be considered cost effective is that it cost less than $50,000 per Quality Adjusted Life-Year.

For men at highest risk (those with a previous fracture), DEXA and bisphosphonate therapy was cost effective. In men older than 80 years, even amongst those without a fracture history, DEXA and bisphosphonate was also cost effective. In men between the ages of 70-80 years, unless the cost of bisphosphonate therapy could be reduced by half, intervention was NOT cost effective (cost would be approximately $100,000/QALY, or double the acceptable threshold of cost-effectiveness).