Recent Posts
Featured Posts

Tainted evidence Part 2

A critical analysis of Rebecca Dekker's "evidence based" circumcision article

October 26, 2019

Rebecca Dekker PhD RN, a nurse and founder of the Evidence Based Birth website, posted an article, "Evidence and Ethics on: Circumcision." [1] An FAQ states that the article provides "evidence-based information as clearly, completely, and objectively as possible... [W]e endeavored to leave our personal opinions out of this article on circumcision and highlight the voices of experts in the medical and bioethics literature." [2]

Guest contributor Melanie Lindwall Schaab MS RN wrote this comprehensive analysis. Schaab is a certified family nurse practitioner, a licensed midwife, and an admin for Circumcision Facts and Science, Dekker's statements are highlighted in blue, followed by Schaab's responses

Tainted evidence Part 1

Urinary Tract Infections (UTIs)

any research on the proposed benefits of circumcision that involves healthy newborns who were circumcised is observational.

She falsely states that all research on the effect of circumcision on the rate of UTIs is observational; therefore she dismisses all of it as not very informative. Actually to my knowledge there is one published RCT [41]

Because observational studies are not randomized, we cannot rule out that other factors may be at least partially responsible for the observed effects of circumcision. For example, preterm infants with low birth weight may be more likely to get a UTI and more likely to be intact (if they were considered too fragile for the circumcision surgery after birth), and this would show an increase in UTIs among intact infants (Van Howe, 2005).

She falsely implies (citing an anti-circumcision activist named Van Howe who is known for his dishonesty in research) that UTI results were biased by the inclusion of premature infants (who are more likely to have UTIs and also more likely to be uncircumcised). However, I would challenge her to provide just one study of circumcision vs. UTIs that included premature infants. There isn't one. (This is a weakness of circumcision research, since it cannot give us a risk-benefit ratio of elective circumcision in preemies for the prevention of UTIs.) So the results were not biased by the inclusion of premature infants.

In Table 1, where she lists the evidence on UTIs as Low quality and Weak, she should have labeled it as High quality and Strong because it includes an RCT and is not biased by the only issue she suggested (preemies). The table specifies RCTs on newborns who have not yet experienced the issue in question. In other words, she phrased it precisely to exclude the RCT from Turkey, because the youngest age was 3 months (not a newborn) and because they had already experienced a UTI. [37] In other words, she phrased it to eliminate any RCT on the subject, so that she could falsely label the evidence on UTIs as low quality and weak.

Ways to reduce risk of UTIs include breastfeeding, healthy bathroom habits, avoid constipation, healthy diet, avoid irritants, and minimizing antibiotic use.

In infancy, formula feeding appears to increase the incidence of UTI by 35%--i.e., formula-fed babies have 1.35x as many UTIs as exclusively breastfed babies [42]. However, uncircumcised infant boys have ten times more UTIs than do circumcised infant boys [43-47]. While I am a huge breastfeeding advocate, I'm not going to give a false impression of its effects in order to coerce or trick parents into doing what I think they should do.

The other preventive measures she recommends don't apply to infants, especially those younger than 6 months who are not yet eating solid foods. Since infancy is when UTIs are most common and most dangerous, that's an important distinction. Her recommendation to minimize antibiotics to prevent UTIs is ironic, since uncircumcised boys are more likely to need antibiotics.

About 111 circumcision would be needed to prevent a single (treatable) UTI in infancy.

Despite the evidence in the study clearly showing a significant benefit of circumcision, the authors reached an anti-circumcision conclusion. [48] As it is the third most recent meta-analysis on the subject, it's strange that she didn't use one of the two more recent meta-analyses. [49] The most recent meta-analysis (2013) reported that about 7% of uncircumcised boys will have a UTI by age 7. They calculated that the number of circumcisions needed to prevent one UTI across the lifespan (NNT) is about 4.6. (While they don't give a NNT for infancy, if 1.5% of uncircumcised boys will experience a UTI in infancy, the NNT for infancy is roughly 70.) They found that the lifetime risk of UTI is 32% for uncircumcised males and 9% for circumcised males. [50] (The female lifetime UTI risk is greater than 60%.) [51]

Phimosis and balanitis (inflammation)

[No] randomized [research] on healthy newborns not yet experiencing these medical problems [because] the research [regarding phimosis and balanitis] is on treatments for boys with physician-diagnosed phimosis.

Her statement is extremely specific to the effect that she rejects any research on the subject. The studies on the proportion of boys in each group who experience overall penile problems including phimosis and infections are not RCTs and don't assess boys who were "not yet experiencing these medical problems" - because nearly all newborns do have phimosis. It's the normal state of the newborn foreskin. So of course they were all affected before their circumcisions.

Her rating of the evidence on phimosis is disingenuous at best. For example, she rates the evidence on topical steroids for treatment of pathological phimosis--in which case, the boy is already experiencing pathological phimosis*--as moderate to high quality. However, she requires the research on circumcision to involve only boys who have NOT already experienced phimosis. Since nearly all boys experience phimosis from birth and since studies of circumcision (or anything else) for *treatment of pathological phimosis involve boys who obviously already have pathological phimosis, this means she (a) phrased the question in such a way that all research on circumcision and phimosis would be excluded and (b) is using different standards for research on circumcision and research on alternatives to circumcision.

She rates the research that was not about circumcised vs. uncircumcised boys on these issues and therefore cites very uninformative and outdated studies. For example, she cites a 1999 study by Dr. Rickwood for the incidence of pathological phimosis in uncircumcised boys, which she puts at 0.6%. However, research has shown that the incidence of pathological phimosis, mostly due to a skin disease called lichen sclerosus (LS) or balanitis xerotica obliterans (BXO), is increasing [52], and so more recent research is necessary to draw an incidence estimate. In fact, another article by the same author she had cited, Dr. Rickwood, published only 3 years later reported the incidence at about 1.5% [53]. Another important point is that the incidence of LS/BXO peaks in males aged in their 30s, which means it's very short-sighted to only look at the incidence in children, and it affects 35-55% of uncircumcised diabetic males [54]. Since about 20% of males will develop diabetes, that means 7-11% of uncircumcised males will develop LS/BXO due to diabetes. This does not include the proportion who will develop LS/BXO in adulthood due to other causes.

It's also important to note that she understates the issue when she implies that the only form of phimosis that may require treatment is caused by LS/BXO. A 2016 study from Denmark (a country very hostile to circumcision) reported that 5.5% of boys required treatment for phimosis. [55]

Her section on balanitis is also very confusing. She cites one study here and one study there (in 1997 and 1999) which confirms her opinion, but does not cite higher quality evidence such as a systematic literature review or meta-analysis. I found two such recent articles which reported that circumcised males experienced a nearly 70% reduction in balanitis [54] and another which did not provide a meta-analysis but reported similar incidences in circumcised and uncircumcised males [56].

The risk of phimosis as a complication following circumcision is also about 1%, which occurs when the scar tissue covers the glans.

It's unclear where she got this idea; that figure contradicts other research on the subject. Such a problem would require re-circumcision, which is only done in about 0.2% of newborn circumcisions (usually for cosmetic reasons). [57] Correctional procedures up to age 18 were performed on 0.3% of circumcised boys. [36] The studies show that phimosis affects 6% of uncircumcised boys and less than 0.2% of circumcised boys - not her figures of 0.6% of uncircumcised boys and 1% of circumcised boys.

Human immunodeficiency virus (HIV)

I've already addressed her HIV arguments, so I'll skip that part of the table.

Human papillomous virus (HPV)

Condoms, HPV vaccine, abstinence, or mutual monogamy with an uninfected partner [are] non-invasive methods of prevention or treatment. Consistent condom use does reduce the risk of HPV. [58] However, as I discussed earlier, many men don't use condoms even when extensively and repeatedly educated about the importance of condoms and given plenty of free condoms. If men did use condoms consistently as we've been advising for decades, we wouldn't have the STD epidemic that we do. The most recent meta-analysis indicates that a vaccine does not prevent HPV infection in males. (It does in females.) In fact, the meta-analysis found a non-significant increased incidence of HPV infection in vaccinated males. [59] Abstinence is great for young people, but realistically most people eventually have sex. Monogamy requires that the individual depend on his partner to practice safe sex. So why not circumcision? Basically, of all the options to reduce HPV risk, circumcision is the best option. It actually works (unlike the vaccine for males). It does not make unrealistic expectations of the male (like condoms and abstinence). It does not rely on the faithfulness of his sexual partner (like monogamy). Because the evidence in favor of the HPV preventive effect comes from high-quality observational research and one of the highest-quality types of study, very large RCTs, the evidence is High quality and Strong, not moderate quality and weak. The most recent meta-analyses all report more than 40% reduced incidence of HPV in circumcised males. [60] [61] [62] [63]

Herpes Simplex Virus (Herpes)

Abstinence or mutual monogamy with an uninfected partner [are] non-invasive methods of prevention or treatment; consistent condom use provides the same (low) level of protection as circumcision.

I addressed abstinence and monogamy above. The most recent systematic review found that condoms are ineffective against herpes. "Using condoms during more than 25% of sex acts was associated with a 92% reduction in the risk of women acquiring HSV-2 but was not associated with a protective effect among men." [64] By contrast, circumcision reduces herpes risk by roughly 30%. [65] Because that data includes RCTs, it is "high quality" and "strong" evidence, not "moderate quality" and "weak." Circumcision is also effective at treating herpes according to research from India, a country which has an anti-circumcision bias. They found that herpes outbreaks were 21 to 25 times higher before circumcision and in men who remained uncircumcised than in men after circumcision. Circumcised men also enjoyed longer disease-free periods between outbreaks. [66]

For reasons unclear, she does not provide information on several STIs and STDs, including genital ulcerative disease, chancroid, lymphogranuloma venereum, granuloma inguinale, chlamydia, gonorrhea, trichomonas, hepatitis B, mycoplasma genitalium, genital warts, non-gonococcal urethritis, scabies, and pubic lice. Besides phimosis she also does not address structural problems with the penis, such as paraphimosis, frenulum breve, buried penis, chordee, pearly penile papules, penile lymphedema, and penile mondor's disease. [67]

Penile Cancer

Most penile cancer is linked to HPV. Researchers today believe that HPV and lichen sclerosus (LS) are the most important risk factors for penile cancer. I already addressed HPV above. She indirectly refers to LS when she says to "treat phimosis that persists into adulthood". LS causes phimosis and there is no increased risk of penile cancer in uncircumcised men without phimosis compared to circumcised men, but there is up to 37 times higher risk of penile cancer in uncircumcised men with phimosis, indicating the strong effect of LS on penile cancer. [68] [69] In fact, the incidence of HPV is also much higher in men with phimosis than in men without phimosis. For example, a 2016 Brazilian (anti-circumcision) study found that HPV affected 47% of phimotic men vs. 16% of non-phimotic uncircumcised men, and that 50% of the HPV samples from phimotic men were high-risk (i.e., cancer-causing) HPV compared to only 1 man without phimosis [70].

The link specifically to LS, not just phimosis, has also been studied; it's now undeniable that LS is a primary, if not the primary, cause of penile cancer. Researchers have not yet found any way to prevent LS aside from circumcision. Treatment typically involves steroids and circumcision or steroids alone, but to my knowledge, there is no research showing that either treatment prevents penile cancer.

According to the AAFP, 300,000 circumcisions would need to be done to prevent one case of penile cancer.

Actually it's closer to 600-900 circumcisions since 1 in 600-800 men not circumcised in infancy will be affected by penile cancer compared to 1 in 50,000-12,000,000 men circumcised in infancy. According to a study that I cited earlier, 1.9 circumcision complications would occur for every 1 case of penile cancer prevented. [50] And since penile cancer is far more serious than the average circumcision complication, that means the benefits outweigh the risks on this one issue alone.

The reason that no RCT has assessed circumcision for the prevention of penile cancer is obvious. Since infant circumcision protects against penile cancer, but circumcision later in life does not, one would have to randomly assign thousands of infants to be circumcised or to remain uncircumcised (again, there's the issue of getting parents to agree to it) and then track them for more than 90 years each. (The average age of penile cancer in a circumcised man is about 88, while the average age in an uncircumcised man is several decades younger) The fact that we don't have an RCT on a subject that is not practical to study via RCT should not count as evidence against this point. It's a similar reason why RCTs were conducted on adult men to study the effect of circumcision on HIV: because if they did infants, they'd have to track them for several decades.

It's puzzling that she doesn't mention prostate cancer, the most common cancer to affect men. In the most recent meta-analysis I could find, researchers found that circumcision reduces prostate cancer by 10% and aggressive prostate cancer by 16%. (Among men with prostate cancer, circumcised men are less likely to have a serious case.) [71]

What are the potential health risks from circumcision?

Complications that can occur during and soon after circumcision

The true rate of complications after newborn circumcision is not known. Part of the reason for this is that researchers are focused on finding the potential benefits from circumcision (e.g., protection from HIV), not the potential harms.

Actually the reason why there's so much research on HIV is because it's a terrible scourge. It is the primary cause of infectious disease deaths in the world, and it spreads rapidly. Most of the early studies on circumcision and HIV were actually studies of everything versus HIV, and coincidentally discovered an association with circumcision. (They first discovered that religious and cultural groups that circumcise had lower rates of HIV than groups that did not circumcise, whereupon they started specifically seeking out circumcision information.) If you look at research on the benefits of, say, PrEP or ART against HIV, you will see a similar pattern with far more studies on efficacy than on safety. Exactly the same pattern occurs with studies on vaccines. For example, year after year, the Cochrane Collaboration has reported that hundreds of studies have assessed the efficacy of the flu vaccine, but none (literally 0) have assessed whether the flu vaccine can trigger the development of pneumonia, in spite of case studies suggesting that effect. That doesn't mean researchers are wholly unconcerned with the safety of PrEP or ART or anything else. It simply means we need to know with absolute certainty whether the intervention in question even works because if it doesn't work, then safety is irrelevant, whereas some risk is always acceptable if the benefits are greater. (Think of it this way: if you had to prove safety first beyond a shadow of a doubt before researching efficacy, no chemo drug would ever be approved.)

In fact, we recently discovered that use of PrEP (pre-exposure prophylaxis, a drug you can take before or a couple days after having sex that reduces your likelihood of catching HIV from that person) reduces the use of condoms--i.e., when PrEP is available, people feel less pressure to use condoms. However, we had already discovered that, whatever the risks, PrEP is effective. Since we already knew it is net effective, the fact that it results in reduced use of condoms was irrelevant to the decision of whether to use it. Research after that discovery focuses on how to reduce the risk that people will stop using condoms after PrEP becomes available in their community (i.e., EFFICACY of intervention to improve condom use), rather than conducting yet more studies on the degree to which people stop using condoms (i.e., SAFETY of PrEP). So the effect remains that there are more studies on efficacy than on safety.

Another reason there are more studies on efficacy of circumcision than on safety is that you can't always study both at the same time like you can with things like vaccines or other drugs. At the same time that you give a drug--say, ART (anti-retroviral therapy against HIV)--you can assess whether they had any side effects (safety) AND whether the viral counts decreased (efficacy). However, with circumcision, you can usually only assess the efficacy and only sometimes the efficacy and the safety. For example, most U.S. males are circumcised in infancy, so you couldn't simultaneously assess their risk of STDs, penile or prostate cancer, penile structural problems, skin diseases, etc., all of which generally occur after infancy. And while you can easily ask a group of circumcised and uncircumcised men about various issues and test them for various problems (e.g., a blood test for HIV), you can't easily recruit several physicians or hospitals that perform circumcisions and get permission to comb through their co