Clinical Briefs

With Comments from Russell H. Greenfield, MD, Dr. Greenfield is Clinical Assistant Professor, School of Medicine, University of North Carolina, Chapel Hill, NC; and Visiting Assistant Professor, University of Arizona, College of Medicine, Tucson, AZ.

Oh, to B12 Again … Metformin Use and B12 Deficiency

Source: Ting RZ, et al. Risk factors of vitamin B12 deficiency in patients receiving metformin. Arch Intern Med 2006;166:1975-1079.

Goal: To identify risk factors that would permit prediction of vitamin B12 deficiency developing in people taking metformin for diabetes (DM).

Study design: Nested case-control study performed in Hong Kong using subjects taken from a source cohort population.

Subjects: Those who for clinical reasons had levels of both hemoglobin A1c and vitamin B12 checked in a central laboratory, and who also had at least one year of continuous medical history available (n = 155 identified with B12 deficiency together with n = 310 controls with normal B12 levels).

Methods: A case was defined as an individual with diabetes (DM) who had vitamin B12 deficiency while being treated with metformin. Once identified, two controls matched to the date of blood sampling were selected for each case. Analysis of results proceeded and was repeated after 113 subjects with borderline B12 levels were excluded.

Results: Each 1 g/d incremental increase in metformin dose conferred an increased risk (odds ratio [OR] = 2.88) for development of B12 deficiency. Post-hoc analysis revealed an increased risk for metformin-associated B12 deficiency with dosages higher than 2.0-3.0 g/d. Those who had used metformin for three or more years also had an increased risk of developing B12 deficiency (OR = 2.39). Increased age was related to metformin-associated B12 deficiency, but not in a clinically significant fashion.

Conclusion: In this observational study, dose of metformin employed and duration of therapy were both associated with an increased risk for development of vitamin B12 deficiency, even after adjustment for potential confounders.

Study strength: Regression analysis repeated with exclusion of borderline cases of vitamin B12 deficiency due to lack of data on methylmalonic acid levels.

Study weaknesses: No assessment of calcium intake; retrospective nature of study; inability to adequately track use of H2 blockers or proton pump inhibitors.

Of note: Patients with pernicious anemia were excluded from the study; serum methylmalonic acid (a metabolite of B12) determinations were not routinely performed, nor were homocysteine levels; impact of potential confounders like use of histamine H2 receptor antagonists or proton pump inhibitors was evaluated (of which there was no impact detected, but data were incomplete); B12 deficiency was found more often in vegetarians (many vegetarians and most vegans are at risk for deficiencies of vitamin B12, folic acid, vitamin D, iron, calcium, and zinc), a finding that reached borderline significance.

We knew that: Metformin increases insulin sensitivity, protects against vascular complications associated with DM, and reduces all-cause mortality in overweight people with DM; early clinical studies reported a 30% prevalence of B12 deficiency in patients treated with metformin; reported metformin-associated decreases in vitamin B12 levels range from 14% to 30%; increasing calcium intake appears to reverse B12 malabsorption associated with metformin use (metformin may impair calcium availability, thus impacting the calcium-dependent process of vitamin B12 absorption); existing data suggest that metformin use does not alter intestinal motility or lead to bacterial overgrowth.

Comments: Many of our patients with diabetes are on metformin (trade name Glucophage®) because of the compelling data on its therapeutic efficacy. As is the case with most any clinical intervention, as experience with a therapy grows so, too, do precautions. Vitamin B12 deficiency is an all-too common and yet under-recognized disorder that can result in or at least worsen depression, dementia, peripheral neuropathy, anemia, fatigue, and constipation. Certain medications, like proton pump inhibitors, have been previously implicated in vitamin B12 deficiency, but relatively few practitioners are aware of this potential side effect. Given the widespread use of metformin, it's important to emphasize the need for screening for B12 deficiency in those at risk based on the results of this trial.

What to do with this article: Keep a copy on your computer.

Living with Syn: Synbiotics and Colon Cancer

Source: Rafter J, et al. Dietary synbiotics reduce cancer risk factors in polypectomized and colon cancer patients. Am J Clin Nutr 2007;85:488-496.

Goal: To assess the effect of a specific synbiotic preparation on the risk of colon cancer in humans.

Design: Phase II randomized, double-blind, placebo-controlled trial over 12 weeks in Ireland.

Subjects: Participants either had known colon cancer (n = 37, data available for assessment on 34) or were status post-polypectomy (n = 43, data available for assessment on 40).

Methods: Subjects were randomly assigned to receive the synbiotic (oligo-fructose-enriched inulin plus Bifidobacterium lactis Bb12 and Lactobacillus rhamnosus GG) or placebo, and were given a sufficient amount of product for six weeks at study onset and again after six weeks. Appointments with the study nurse took place prior to any intervention, and at six and 12 weeks. Blood and fecal samples were obtained at each visit, and sigmoidoscopy with biopsy samples was performed both before synbiotic administration and at 12 weeks. A comprehensive battery of biomarkers was employed and included immunologic, inflammatory, and genotoxic markers. Subjects maintained diaries during each six-week phase.

Results: Administration of the synbiotic created significant changes in the fecal flora, with increases in bifidobacteria and lactobacilli, and in polypectomized subjects a decrease in Clostridium perfringens. Synbiotic administration resulted in a significantly decreased level of DNA damage in the colonic mucosa and, in participants with polyps, reduced colorectal proliferation, improved barrier function, and decreased cytotoxicity of luminal contents. There was no impact of synbiotic administration on apoptosis, tumor cell invasion, or colonic inflammation.

Conclusion: Specific biomarkers of colon cancer may be favorably altered through administration of a specific synbiotic preparation.

Study strengths: Evaluation of both people with colon cancer and those at high risk for development of colon cancer; tests performed to assess probiotic survival after intestinal transit.

Study weaknesses: Short duration of intervention; small sample size.

Of note: Participants had to have undergone resection for confirmed colon cancer within the prior five years, or have histologically confirmed adenomatous polyps; intestinal bacteria can produce genotoxic, carcinogenic, and tumor-promoting substances from ingested foods; germ-free rats have a lower incidence of chemically induced tumors than do rats with an intact microflora; subjects with known lactose intolerance were excluded from the trial, as were users of non-steroidal anti-inflammatory agents and antibiotics; no adverse effects were reported in association with synbiotic administration during the study.

We knew that: Probiotics have been defined as live microbial food ingredients that may be beneficial to the health of the host; prebiotics have been defined as nondigestible food ingredients that selectively stimulate growth of particular bacteria in a manner beneficial to the host; synbiotics are combinations of prebiotics and probiotics; animal data using prebiotics and probiotics are strongly suggestive of a chemopreventive effect against colon cancer, but there is a paucity of supportive data in humans; mortality from colorectal cancer is second only to lung cancer in men, second only to breast cancer in women; the balance of microbial types in the gastrointestinal tract appears to play an important role in risk for development of colorectal cancer (bifidobacteria and lactobacilli have less propensity to enzymatically produce carcinogens than do bacteria like Bacteroides and clostridia); prebiotics increase butyrate production, which inhibits survival of colon cancer cells and also helps normal colon cells survive.

Clinical import: As the authors point out, research results suggest that diet plays a role in the risk for development of colorectal cancer, meaning that risk may well be modifiable through dietary modification. Add to this the compelling concept that modification of one's intestinal microbiota through oral administration of live microbes might likewise have a chemopreventive impact. The data presented in this article pique our interest significantly, but are not definitive as questions remain. Why did subjects with colon cancer respond differently to synbiotic therapy than those with polyps? What would a trial of longer duration have shown? Would similar results occur if a functional food source of synbiotic were employed? The good news is that a large body of research suggests both prebiotic and probiotic therapy (with far more data available regarding the latter) are safe. In lieu of further data, practitioners can feel comfortable with patient recommendations that include adherence to a healthy diet, and perhaps even taking a daily dose of bacteria.

What to do with this article: Keep a hard copy in your file cabinet.

Pressure and Pain—Analgesic Use and Blood Pressure

Source: Forman JP, et al. Frequency of analgesic use and risk of hypertension among men. Arch Intern Med 2007;167:394-399.

Goal: To assess the association between frequency of non-narcotic analgesic use and incident hypertension over a period of four years.

Design: Prospective cohort analysis.

Subjects: Male health professionals (n = 16,031 with a mean age of 65 years in 2000) without a history of high blood pressure at baseline (part of the Health Professionals Follow-up Study).

Methods: Subjects regularly completed biennial questionnaires about health-related behavior and medical events as part of the long-term study, with food-frequency questionnaires filled out every four years (data from 1998 were imputed for 2000). Detailed information about analgesic use was first obtained in 2000 and included questions about frequency of use in days as well as number of pills consumed per week. Data regarding other parameters such as body mass index (BMI), age, physical activity, and smoking status were obtained in 2000 and again in 2002.

Results: A total of 1,968 new cases of hypertension were reported. An independent association between frequency of analgesic use and incident hypertension was found for all three analgesic classes. Men who used acetaminophen 6-7 days/week had a relative risk (RR) for incident hypertension of 1.34 after adjustment for potential confounders. Employing similar analyses, users of nonsteroidal anti-inflammatory agents (NSAIDs) and aspirin likewise were found to be at increased risk (RR = 1.38 and 1.26, respectively). Secondary analyses showed similar results. Significant findings were also found for acetaminophen and aspirin for number of pills per week and risk for subsequent hypertension. While BMI had no identifiable impact on aspirin use and risk of hypertension, the association between use of acetaminophen and risk of hypertension was higher in men with BMI < 25 kg/m2, while that for NSAID frequency and risk of incident hypertension was higher for men with BMI > 25 kg/m2. A history of smoking was consistently more common with increasing use of non-narcotic analgesics.

Conclusion: Frequency of non-narcotic analgesic use is independently associated with a moderately increased risk for developing of hypertension.

Study strengths: Statistical models employed that accounted for factors like race, age, and family history; additional analyses performed that investigated the potential confounding impact of age on incident hypertension.

Study weaknesses: Dosage information only queried for aspirin; no information on past use of analgesics.

Of note: Non-narcotic analgesics may increase blood pressure through a variety of mechanisms including inhibition of vasodilatory prostaglandins, impaired endothelial dysfunction, enhanced renal sodium and water reabsorption (NSAIDs), and increased cellular oxidative stress (acetaminophen); in this study, men who were using antihypertensive agents without a history of hypertension were excluded from participation; some studies suggest that NSAIDs may antagonize the effects of blood pressure medications; the Health Professionals Follow-up Study began in 1986 with 51,529 participants.

We knew that: Non-narcotic analgesics (acetaminophen, NSAIDs, and aspirin) are the most commonly used drugs in the United States; an association between frequency of analgesic use (measured in days/month) and risk of later development of hypertension was identified in two large cohort studies of women; acetaminophen inhibits prostaglandin H2 synthase, the same enzyme targeted by aspirin and NSAIDs but at its peroxidase, not cyclooxygenase, catalytic site, creating differential tissue specificity; higher BMI is associated with endothelial dysfunction, salt sensitivity, and oxidative stress.

Clinical import: The sheer number of people who use over-the-counter pain medication and deem it benign is worth pause, as is the sobering incidence of cardiovascular disease in the United States. Is it possible that the two could be related? The results of this trial would suggest so, and raise the specter of a significant public health issue. Before jumping to conclusions, however, it's important to recognize that frequent use of analgesics can be a marker for other behaviors that may also contribute to development of hypertension (as noted in this study with regard to the prevalence of smoking among users of analgesics). The results also beg the question of whether specific complementary therapies might obviate such risk. Could herbal anti-inflammatory agents like turmeric lessen discomfort without increasing the risk for hypertension? If acupuncture were an accessible and reimbursable treatment for headaches, would the cost:benefit ratio be all the more intriguing in light of these data? We await further clarification, but the importance of emphasizing to our patients that non-narcotic analgesics do have potential side effects is now all the more evident.

What to do with this article: Keep a copy in your file cabinet.