Circumcision and its alternatives in Germany: an analysis of nationwide hospital routine data

A recent study demonstrated an increasing frequency of neonatal circumcisions in the United States following publication of the most recent American Academy of Paediatrics guideline on the subject from 2012, which expresses an affirmative stance on neonatal circumcision after more than a decade of equipoise [2]. Interestingly, a similar trend has been described even before the millennium using routine hospital data [21]. In preceding studies, two moderator variables have been identified: Medicaid coverage of circumcision was linked to a 24% higher rate of neonatal circumcision compared to non-Medicaid-coverage, and Hispanic descent was associated to lower circumcision rates [22]. Consequently, the number of neonatal circumcisions dropped in states that withdrew Medicaid coverage for neonatal circumcisions, but remained stable in states that did not [23]. Likewise, this study also demonstrated a reduced frequency of neonatal circumcision among children of Hispanic descent and a slightly reduced frequency of neonatal circumcisions among non-Hispanic black children [23]. For Germany, circumcision has been associated to being born into a migrant family, particularly being descended from Turkish families with a more than tripled rate of circumcisions compared to the rest of Germany’s paediatric population [24]. Many et al. [2] used the Paediatric Health Information System database that covers free-standing children’s hospitals from many states from the United States of America, although particularly those in metropolitan areas, and has been described to be somewhat representative for the whole paediatric population in the United States of America [25]. Of note, the moderator variables were associated with a reduced frequency of neonatal circumcisions, but not an increase, which has not been described for any other potential factor [22, 23].

This is in contrast to other countries such as Australia for which a decreasing rate of circumcision has been described in the same time period [26] and beyond [27]. Similar trends were observed in England [28] and Northern Ireland [29], but large-scale population-based data is scarce and reasons remain unclear. Using nationwide administrative data for hospital reimbursement, we were also able to show a decreasing trend in circumcisions in German hospital-based procedures among all age groups. However, the preputium-preserving operations also showed a declining trend in all age groups, except those in their first year of life. This could be linked to the general reduction in surgical diseases described before, for example in infantile hypertrophic pyloric stenosis [30] or inguinal hernia [31]. An alternative explanation might be a shift of these operations towards outpatient surgery: The vast majority of paediatric surgical departments offer circumcision as an outpatient procedure [32], but this is unlikely as these procedures have been conducted primarily on outpatients even back in the 1980s [33] as it was the case in the United States, too [34]. Although there is no universally accepted age threshold that would preclude outpatient surgery, in Germany, it is uncommon below the age of six months. An individual decision between caretakers, surgeons, and anaesthetists should be reached on whether the procedure can be performed as an outpatient surgery [35]. Consequently, we are confident that the present data covers all operations in neonates, except ritual circumcisions performed by non-physicians, but there is considerable uncertainness in older age groups. This is supported by a cohort of boys that received a preputial operation from an office-based paediatric surgeon, in which none of the patients were operated on in their first year of life [36].

The obvious difference of our data to those by Many et al. is the age distribution between the two cohorts: They reported 67% of all circumcisions in the first year of life, whereas only 12.5% of circumcisions in our cohort were performed in this age group. For the preputium-preserving operations, this number was even smaller with 7.2% of all procedures. This difference is not surprising as the German guideline only recommends circumcision for Lichen sclerosus and high-grade urologic malformations in order to prevent recurrent infections, which leaves only phimosis as a possible indication for surgery and only after a local therapy with steroids has been applied for a sufficient amount of time [37]. Similarly to the observation by Many et al. [2] for circumcisions, the effects of guideline changes could also be observed in Germany, but the other way round: The first step of a local treatment with steroids [37] resulted in a directly observable change in therapy preferences. In 2005, local steroids were never used, but became the definitive treatment for the majority of patients in a single-centre analysis from a tertiary paediatric surgery unit [38]. A result that has been validated in later cohorts [39] and thus reflects the different recommendations made by the German guidelines compared to the guideline by the American Academy of Paediatrics: An affirmative stance towards neonatal circumcision [40] versus primary non-operative treatment that considers a primary circumcision without preceding local steroid therapy in the absence of a balanitis xerotica obliterans or high-grade urinary tract anomalies as not lege artis [41]. For the United States of America, there is no recommendation by the American Urological Association towards the use of topical steroids for phimosis [42]. However, results from a cohort treated by an office-based paediatric surgeon before 2015 showed that only 11% of operated boys received preceding treatment with local steroids [36], so guideline compliance may vary between hospital- and office-based surgeons. On the contrary, in a cohort from the Capital region of Denmark, the majority of patients received topical steroids as first-choice treatment of phimosis and had a similar age-distribution as in the our report using administrative data, too [43].

Although ritual circumcisions should in theory not be conducted within the health-care system, it is likely that they are present [44]. Their relevance has been quantified in data from England based on the comparison with the expected incidence of phimosis [28]. Based on the report that in a German cohort of 176 boys treated for phimosis by an office-based paediatric surgeon, the decision for treatment was based on a non-retractable preputium [36], and taking into account that a non-retractable foreskin is present in 50% of first-graders [45], it is unclear how many circumcisions were performed for therapeutic reasons. Non-medically indicated circumcision in boys in Germany can be performed explicitly by discretion of parental custody since 2013. In the first six months of life, circumcision may be performed by non-physicians if this person is designated to do so by a religious community [46]. Consequently, the latter cases will not be represented within our data if there is no immediate major complication that required therapy on an inpatient basis. Currently, there is no estimate on the number of these procedures conducted by non-physicians without adequate anaesthesia [47]. This might be a general issue due to, perceived, profanity of topics like circumcisions, exemplified by unpublished theses with relatively large cohorts examining the efficacy of local steroids [38, 39], whereas small cases series with single digit numbers of patients still easily appear in the international literature if they present something novel or deal with more prestigious diseases [48]. This is important in so far, as truly representative cohort studies found a more than doubled prevalence of circumcision in boys from families with a migrant background and even a more than tripled prevalence in those from Turkish families [24]. This suggests that ritual circumcisions frequently occur.

A major limitation of our data is its sole focus on hospital-based procedures. As described before, circumcision and its alternatives are often conducted on outpatients [33, 34] and office procedures are recompensed to the surgeon via a different way than it is the case for hospitals. Consequently, our data does not offer the full picture of circumcision and its preputium-preserving alternatives as the distribution of procedures may be different in outpatients. The generalisability to other countries that do not endorse neonatal circumcision is therefore unclear due to the different structure of the German health system that favours inpatient-based approaches due to highly different reimbursements compared to office-based procedures [49]. Another limitation is the focus of administrative data used by us on cases instead of patients: The same patient treated with a preputium-preserving procedure might later on reappear in the data with a circumcision, but this cannot be tracked due to the case-based approach of the data. It may thus also be possible that there is no reduction in foreskin surgery, but just a shift towards outpatient-clinics or office-based surgeons, because these procedures cannot be tracked by the administrative data used by us. This limitation would not apply to databases of the statutory health insurances, because they are patient-based instead of case-based. Other limitations are those that are inherent in the use of secondary administrative datasets: Lack of clinical and demographic details due to the focus on reimbursement of hospitals, misclassifications due to coding errors, and systematic errors introduced by variations in coding practice, because the coding is done at the local hospital level.

Other factors such as information of parents on the altered risk of sexually transmitted diseases following circumcision are unlikely to have effects on parental decisions for or against circumcision [50]. Likewise, we also consider potential effects of a vaccination against human papilloma virus diminishable, because this vaccination is only among the lists of services covered from 2019 onwards by the statutory health insurances in Germany that serve around 90% of Germany’s population [51].

However, our study offers a first comparator to the study by Many et al. [2] from a country that does not endorse neonatal circumcision despite its limitations. Moreover, it might serve as a vanguard to prompt others to contribute population-based data on circumcision and its preputium-preserving alternatives as it has been done for infantile hypertrophic pyloric stenosis [30] or inguinal hernia [52].