Antibiotic Consumption Patterns in European Countries May Be Associated with the Incidence of Major Carcinomas

The possible role of the altered intestinal microbiome in the development of malignancies has been raised recently in several publications. Among external factors, antibiotics are considered to be the most important agent capable of producing dysbiosis in the gut flora, either temporally or permanently. The human microbiome has several beneficial effects in terms of maintaining appropriate human health, but its alteration has been implicated in the development of many illnesses. Our basic aim was to explore a possible relationship between the consumption of different antibiotic classes and the incidence of the most common cancer types (male, female) in European countries. A database of the average, yearly antibiotic consumption (1997–2018) has been developed and the consumption figures were compared to the eight, most frequent cancer incidence calculated for 2018 in 30 European countries. Pearson correlation has indicated different degrees of positive (supportive) and negative (inhibitor) significant associations between antibiotic consumption figures and cancer prevalence. It has been observed that certain antibiotic classes with positive correlation probably augment the incidence of certain cancer types, while others, with negative correlation, may show some inhibitory effect. The relatively higher or lower consumption pattern of different classes of antibiotics could be related to certain cancer prevalence figures in different European countries. Our results indicated that countries with relatively high consumption of narrow-spectrum penicillin (J01CE, J01CF) and tetracycline (J01A), like certain Scandinavian countries, showed a higher incidence of female colorectal cancer, female lung cancer, melanoma, breast, prostate and uterus corpus cancer. Countries with relatively higher consumption of broad-spectrum penicillin (J01CA, J01CR) and some broad-spectrum antibiotics (J01D, J01F, J01M), like Greece, Hungary, Slovakia, France, etc. showed a higher incidence rate of male lung cancer and male bladder cancer. The higher incidence rate of different cancer types showed association with the higher consumption of antibiotics with “augmenting” properties and with less consumption of antibiotics with “inhibitory” properties.


Introduction
The first known cancer cases in humans have been verified in intact mummies from the pharaonic necropolis of Qubbet el-Hawa in Aswan, Egypt (BC 2000-1800), when CT scan was performed on the bodies. Researchers identified breast cancer and multiple myeloma [1].
As of now, 3.9 million new cancer cases and 1.9 million cancer deaths were estimated in Europe in 2018. Cancers of the female breast (523,000 new cases, 13% of all cancer cases), colorectum (500,000, 13%), lung (470,000, 12%) and prostate (450,000, 12%) were the most common cancers on the continent and combined they represented almost half of the overall cancer burden [2].
The transformation from a normal cell into a tumor cell is a multistage process in which growths often invade surrounding tissues and can metastasize to distant sites. Apart from the genetic background, several external agents with carcinogenetic properties, such as chemicals, toxins, irradiations, infections, etc. can contribute to the process. Recently, the possible role of an altered intestinal microbiome has been raised in the process of carcinogenesis by several researchers and the possible effect of certain antibiotics has been mentioned [3][4][5][6][7][8][9]. Tumor-promoting effects of the microbiota in colorectal cancer (CRC) seem to be caused by altered host-microbiota interactions and by dysbiosis, rather than by infections with specific pathogens. Accordingly, germ-free status and treatment with wide-spectrum antibiotics led to a significant reduction of the numbers of tumors in chemical and genetic experimental models of colorectal carcinogenesis. Hence, the strong microbiome-modification capability of antibiotics and their indirect role in the development of malignancies should be considered [10].
It has been reported that increased exposure to antibiotics during 15 years was associated with a significant increase in prostate cancer risk. This association was dose-dependent [11]. For gastrointestinal malignancies, the use of penicillin was associated with an elevated risk of esophageal, gastric and pancreatic cancers [12]. Similarly, a dose-dependent increase in breast cancer risk was observed in association with antibiotic exposure up to 15 years in the past [13]. Accumulating evidence from animal models suggests that specific microbes and microbial dysbiosis can potentiate hepatobiliary-pancreatic tumor development by damaging DNA, activating oncogenic signaling pathways and producing tumor-promoting metabolites [14].

Working Hypotheses/Concept
Based on the extensive documentation of the possible role of the altered microbiome in the carcinogenesis, it may be concluded that agents, like antibiotics, having strong potency of producing alteration of the microbiome, could trigger certain pathologic processes, leading to the development of different malignancies. Substantial variation of cancer incidence and mortality rates are observed at the national level across EU countries and it is logical to suspect that similar variations of the causative agents, including the alteration of the microbiome, may be the driving force behind this phenomenon. Considering the possible role of microbiome-triggered mechanisms in carcinogenesis, we hypothesize that the consumption of different antibiotics, generating different modifications of the microbiome, can contribute to the process of carcinogenesis. The hypothesis suspects the association of antibiotic consumption patterns and cancer morbidity data (prevalence, incidence) in 30 European countries included in the study.
Antibiotic consumption database for comparison has been extracted from the ECDC yearly reports on antibiotic consumption for the years of 1997 to 2018 (22 years) reported from 30 EU countries included in the study [17]. The amount of antibiotic consumption appeared as defined daily dose (DDD) per 1000 inhabitants per day (DID) in the respective countries. Average yearly antibiotic consumptions were calculated for 1997-2018 (22 years) in DID as the total systemic antibiotic consumption (J01) and at ATC (Anatomic Therapeutic Chemical classification) [14] classification Level 3 for the major classes of antibiotics as J01A (tetracycline), J01C (penicillin), J01D (cephalosporin), J01F (macrolide), J01 M (quinolone) and at ATC Level 4 for the narrow-spectrum, penicillinase-sensitive penicillin (J01CE), penicillinase-resistant narrow-spectrum penicillin (J01CF), broad-spectrum beta-lactamase sensitive penicillin (J01CA) and broad-spectrum combined with beta-lactamase inhibitors (J01CR). The total systemic antibiotic consumption (J01) expressed in DID/countries were considered as 100% of the respective antibiotic consumption in the countries included in the study and the amount of antibiotic consumption of J01A, J01C, J01D, J01F, J01M, has been calculated as relative share of the total amount (J01) and expressed in percentage (%). Similarly, all the subgroups of penicillin (J01CE, J01CF, J01CA and J01CR) at ATC Level 4 has been calculated as relative share of the total consumption (J01) and expressed in percentage. Groups of narrow (J01CE+CF) and broad (J01CA+CR)-spectrum penicillin were formulated and featured as cumulative relative share of the total amount of systemic consumption ( Table 2). Rank order (decreasing) with the highest incidence of different cancer types (first 10 countries) has been compared to the rank order of antibiotic consumption data (decreasing) to observe the similarities between the higher cancer incidence and the higher consumption figures of antibiotics probably facilitating the development of cancers. Invers (increasing) rank order of antibiotics with supposedly inhibitor effect on the development of cancer was similarly compared to the rank order of cancer incidence (Tables 3-8). Table 2. Average, yearly antibiotic consumption in 30 European countries for 1997-2018 (22 years) derived from the ECDC yearly reports. Total antibiotic consumption for systemic use (J01), expressed as defined daily dose (DDD)/1000 inhabitants/day (DID). Other classes of antibiotics at ATC Level 3 and 4 (for the penicillin group/J01C/) consumed has been featured as the relative share of the total consumption and estimated in percentage (%).  ATC codes: J01A-tetracycline; J01C-penicillin; J01CA-broad-spectrum, beta-lactamase sensitive penicillin; J01CR-broad-spectrum penicillin combined with beta-lactamase inhibitors; J01CE-narrow-spectrum, beta-lactamase sensitive penicillin; J01CF-narrow-spectrum, beta-lactamase-resistant penicillin; J01D-cephalosporin; J01F-macrolide; J01M-quinolone; J01 B/N-ratio of broad-spectrum and narrow-spectrum antibiotics. Table 3. Rank order (decreasing) of the incidence of male lung cancer (10 countries) compared to the rank order (decreasing) of antibiotics with suspected "promoter" effect and the increasing (reverse) rank order of the suspected "inhibitor" effect of antibiotics. Of the first ten countries with the highest incidence rate of male lung cancer, nine countries can be found among the highest consumption figures of "promoter" antibiotics (exception: Poland). Seven countries with the lowest consumption figures of "inhibitor" antibiotics, are among the first high incidence rate of male lung cancer (exception: Hungary, Lithuania, Estonia) J01CA-broad-spectrum, beta-lactamase sensitive penicillin; J01R-broad-spectrum penicillin with beta-lactamase inhibitors; J01D-cephalosporin; J01F-macrolide; J01CE-narrow-spectrum, beta-lactamase-sensitive penicillin; J01CF-narrow-spectrum, beta-lactamase-resistant penicillin; J01A-tetracycline. Table 4. Rank order (decreasing) of the incidence of female lung cancer (10 countries) compared to the rank order (decreasing) of antibiotics with suspected "promoter" effect and the increasing (reverse) rank order of suspected "inhibitor" effect of antibiotics. Out of the first ten countries with the highest incidence rate of female lung cancer, five countries can be found among the highest consumption figures of "promoter" antibiotics (bold, italics). Eight countries with the lowest consumption figures of "inhibitor" antibiotics are among the firs high incidence rate of female lung cancer (bold, italics). J01CE-narrow-spectrum, beta-lactamase-sensitive penicillin; J01CF-narrow-spectrum, beta-lactamase-resistant penicillin; J01CA-broad-spectrum penicillin; beta-lactamase sensitive penicillin; J01CR-broad-spectrum penicillin with beta-lactamase inhibitors; J01D-cephalosporin; J01M-quinolone. Table 5. Rank order (decreasing) of the incidence of breast cancer (10 countries) compared to the rank order (decreasing) of antibiotics with suspected "promoter" effect. Of the first ten countries with the highest incidence rate of breast cancer, seven countries can be found among the highest consumption figures of "promoter" antibiotics (bold. italics). J01CF-narrow-spectrum, beta-lactamase-resistant penicillin; J01A-tetracycline. Table 6. Rank order (decreasing) of the incidence of prostate cancer (10 countries) compared to the rank order (decreasing) of antibiotics with suspected "promoter" effect. Of the first ten countries with the highest incidence rate of prostate cancer, six countries can be found among the highest consumption figures of "promoter" antibiotics (bold. italics). Similarly, seven countries with the lowest consumption figures of "inhibitor" antibiotics are among the firs high incidence rates of prostate cancer (bold. italics). J01CE-narrow-spectrum, beta-lactamase-sensitive penicillin; J01CF-narrow-spectrum, beta-lactamase-resistant penicillin; J01CA-broad-spectrum, beta-lactamase sensitive penicillin; J01CR-broad-spectrum penicillin with beta-lactamase inhibitors; J01D-cephalosporin; J01M-quinolone. Table 7. Rank order of melanoma in males (decreasing) compared to the rank order (decreasing) of antibiotic consumptions of "promoting" and the rank order (increasing) of "inhibitor" effect of antibiotics. Eight countries with the highest consumption of "promoting" antibiotics included in the first ten highest incidences of melanoma (males) countries. Eight of ten countries with the lowest consumption of "inhibitory" antibiotics are included in the first ten countries with the highest incidence of melanoma (males).  J01A-tetracycline; J01CE-narrow-spectrum, beta-lactamase sensitive penicillin; J01CF-narrow-spectrum, beta-lactamase-resistant penicillin; J01CA-broad-spectrum, beta-lactamase sensitive penicillin; J01CR-broad-spectrum, beta-lactamase inhibitor penicillin; J01D-cephalosporin; J01M-quinolone. Table 8. Rank order of melanoma in females (decreasing) compared to the rank order (decreasing) of antibiotic consumptions of "promoting" effect and the rank order (increasing) with "inhibitor" effect of antibiotics. Eight countries with the highest consumption of "promoting" antibiotics included in the first ten highest incidences of female melanoma countries. Nine of ten countries with the lowest consumption of "inhibitory" antibiotics are included in the first ten countries with the highest incidence of melanoma (females). J01A-tetracycline; J01CE-narrow-spectrum, beta-lactamase sensitive penicillin; J01CF-narrow-spectrum, beta-lactamase-resistant penicillin; J01CA-broad-spectrum, beta-lactamase sensitive penicillin; J01CR-broad-spectrum, beta-lactamase inhibitor penicillin; J01D-cephalosporin; J01M-quinolone.

Statistics
Pearson's correlation was applied for calculating correlation and statistical significance. Positive correlation and significance were estimated when the p value was ≤0.05 and the r was a positive number. Negative (inverse) significance was estimated when the p value was ≤0.05 and the r was a negative number.

Results
All types of cancers included in the study showed significant or statistically insignificant, positive or negative associations with at least one, ore more classes of antibiotics, including with the total consumption (J01) or with the high consumption rate of broad-spectrum antibiotics (J01 B/N). It is of importance that the male and female cancer patients within the same location (colorectal, lung, melanoma, kidney, bladder) in certain cases, showed differently, sometimes opposite, associations with the same antibiotic groups.

Colorectal Cancer
Male patients showed a statistically insignificant negative correlation with tetracycline consumption (J01A), and no other associations were observed. In female cases, the statistically insignificant, negative association was found with the total consumption of systemic antibiotics (J01) and quinolone (J01M) consumption and significant negative (inhibitor) association with cephalosporin (J01D). Strong, positive, supportive, significance was found with the consumption of narrow-spectrum penicillin (J01CE, J01CF).

Melanoma
The statistical analyses resulted in several, positive and negative associations; it did not show any difference between the male and female cases. Melanoma incidence was inversely associated with the total consumption of antibiotics (J01) in both sexes and protective (inhibitor) association was detected between broad-spectrum penicillin combined with a beta-lactamase inhibitor (J01CR), but not with broad-spectrum, beta-lactamase sensitive penicillin (J01CA), which raises the possible protective effect of beta-lactamase inhibitors, most likely clavulanic acid, on the development of melanoma. Similarly, the inhibitor effect was associated with cephalosporin (J01D) and quinolone (J01M) consumption. Strong supportive significance was recorded between melanoma incidence and the consumption of narrow-spectrum penicillin (J01CE, J01CF).

Breast
An only a weak positive correlation was observed with the narrow-spectrum, beta-lactamase-resistant penicillin (J01CF, p = 0.059).

Prostate
Positive significance was seen with the narrow-spectrum, beta-lactamase-resistant penicillin (J01CF). A statistically insignificant similar association was observed when J01CE and CF were calculated together. Negative significance was found with cephalosporin (J01D) and quinolone (J01M).

Uterus Corpus
A weak, negative association was observed with the narrow-spectrum, beta-lactamase-resistant penicillin (J01CF).

Kidney
Inverse, negative, significance was detected between total antibiotic consumption (J01) and the incidence of kidney cancer in both sexes. A weak, positive correlation was found between broad-spectrum, beta-lactamase sensitive penicillin (J01CA) and negative with narrow-spectrum, beta-lactamase-resistant penicillin (J01CF) in male, kidney cancer patients. Female patients showed a statistically insignificant, negative association with the consumption of quinolone (J01M) and broad-spectrum antibiotics.

Bladder
A positive, significant, correlation was recorded with the joint consumption of broad-spectrum penicillin (J01CA + J01CR), but none, with the separate groups (J01CA and J01CR separately). A statistically insignificant, positive correlation was found with the consumption of cephalosporin (J01D) and quinolone (J01M). No, any association was observed in female cases.

Discussion
Gut microbiota is composed of different bacteria species taxonomically classified by genus, family, order and phyla. Gut microbiota consists of not only bacteria, but also viruses, fungi and Archaea. Each individual is provided with a unique gut microbiota profile that plays many specific functions in host nutrient metabolism, maintenance of structural integrity of the gut mucosal barrier, immunomodulation and protection against pathogens, etc. [19,20].
Gut microbiota is shaped in early life as their composition depends on infant transitions (birth gestational date, type of delivery, methods of milk feeding, weaning period) and external factors such as antibiotic use. This personal and healthy core native microbiota remains relatively stable in adulthood, but differs between individuals due to enterotypes, body mass index (BMI) level, exercise frequency, lifestyle and cultural and dietary habits and by gender [19,20].
Dysbiosis or disruption of the normal human microbiota is associated with a wide range of diseases, including inflammatory bowel disease, multiple sclerosis, obesity, autism, depression, cardiovascular disease and allergy, as well as cancer [21][22][23].
The microbiome has been implicated in cancer in a variety of specific ways, including being directly oncogenic, through the promotion of oncogenic mucosal inflammation or systemic metabolic/immune dysregulation and through modulation of anti-cancer immunity or the efficacy of anticancer therapy. Bacterial species are found in tumor tissue itself, normal tissue adjacent to the tumor and at tumor sites such as the gut, genitourinary tract and airway, with overlap between these sites [9]. The highest risk was found in individuals with a long duration of antibiotic exposure or those receiving higher doses. There was a 30% increased incidence of lung, hematological, pancreatic and genitourinary cancers compared to controls due to increased antibiotic exposure [5]. An overall increase of 18% for all malignancies [5]. An extensive meta-analysis showed evidence that antibiotic use slightly increases the risk of hematological (multiple myeloma and lymphoma), gastrointestinal (colorectal, hepatobiliary, pancreatic and gastric cancers), lung and genitourinary cancers (prostate, bladder and kidney). Weak evidence supported the increased risk for breast and other cancers such as gynecological cancers and melanoma. Moderate evidence was found that this risk is associated with specific classes of antibiotics (macrolides, beta-lactams, quinolones, sulfonamides and cephalosporins), but low or insufficient evidence of associations with the other analyzed classes [5]. The extensive use of antibiotics may predict the development of cancer [24]. Even maternal antibiotic exposure showed an association with cancer morbidity in children [25]. Our observations (Table 9.) indicated the possible role of different antibiotics in the development of certain malignancies, probably through the induction of dysbiosis in the gut flora or by modifying the composition of tissue bacteria. Sex differences, observed in our analyses, may be associated with the gender-related difference of the gut flora [26]. Although the microbiome influences carcinogenesis through mechanisms independent of inflammation and immune system, the most recognizable link is between the microbiome and cancer via the immune system, as the resident microbiota plays an essential role in activating, training and modulating the host immune response. In certain cases, mechanisms that are more detailed were observed. The interaction between F. nucleatum Fap2 protein and host polysaccharide (Gal-GalNAc) mediates F. nucleatum colonization in colorectal cancer. F. nucleatum mediates tumor-immune evasion via the T-cell immunoreceptor with Ig and ITIM domains (TIGIT). The Fap2 protein secreted by F. nucleatum interacts with TIGIT and inhibits natural killer (NK) cell-mediated immunosurveillance of cancer [27]. Our concept is further supported by the fact that the rank order (decreasing) of cancer prevalence in countries included in the study and the rank order (decreasing) of the consumption of different antibiotic classes showing positive correlation with the cancer incidence is very similar. The inverse rank order of antibiotics (increasing), showing negative (inhibitor) correlation related to the development of certain cancer types, strengthens the possibility that the less consumption of "inhibitor" antibiotics may increase the incidence of certain malignancies (Tables 3-8).

Conclusions
Our findings strongly support the observations of the role of antibiotics in the development of various malignancies, probably acting through the modification of microbiome and hence, the dominant antibiotic consumption patterns in different countries are reflected in the cancer prevalence data of the given country. Countries with relatively high consumption of narrow-spectrum penicillin (J01CE, J01CF) and tetracycline (J01A), like certain Scandinavian countries, showed a higher incidence of female colorectal cancer, female lung cancer, melanoma, breast, prostate, uterus corpus. Countries with relatively higher consumption of broad-spectrum penicillin (J01CA, J01CR) and some broad-spectrum antibiotics (J01D, J01F, J01M), like Greece, Hungary, Slovakia, France, etc. showed a higher incidence rate of male lung cancer and male bladder cancer. Certain cancers did not show any significance with any classes of antibiotics, like colorectal cancer of males and bladder cancer in females.
Our study included the eight most common cancers in males and females, but further analyses may uncover other possible associations between carcinoma incidence and antibiotic consumption.

Weakness of the Study
The positive and negative correlation between cancer incidence and antibiotic consumption data could not be applied to the individual level and no other confounding circumstances could be identified, which may influence the results.

Strengths of the Study
The positive and negative correlation between the large databases of cancer prevalence and antibiotic consumption strongly supports the role of antibiotics in carcinogenesis as described in the literature. The rank order of cancer incidence in different countries are similar to the rank order of antibiotic consumptions. The higher incidence of different cancers shows higher consumptions patterns of antibiotics with "promoting" effect and lower consumption patterns of antibiotics with "inhibitory" effects on cancer incidence.