We collected data on all incident cases of anogenital cancer in Granada diagnosed between 1985 and 2017. The following ICD-O-3 cancer sites were considered: anus (C21), vulva (C51), vagina (C52), cervix uteri (C53), and penis (C60).

Incidence data were obtained from the Granada Cancer Registry (GCR), a population-based cancer registry in Southern Spain that began its activity in 1985, currently covering a population of 930,000 inhabitants (50.5% women). GCR information sources include clinical records from public and private hospitals, files from Anatomical Pathology and Oncology units, and death certificates. Data from the GCR are regularly published in the IARC monographs “Cancer Incidence in Five Continents” (25). The GCR is a member of the Spanish Network of Cancer Registries (REDECAN) and the European Network of Cancer Registries (ENCR). It collaborates in the international projects EUROCARE (http://www.eurocare.it) and CONCORD (https://csg.lshtm.ac.uk/research/themes/concord-programme/). In terms of data quality, 96.8% of cancers in the anogenital area had pathological confirmation (by histology or cytology) and the death certificates constituted the sole information source in only 0.7% of cases.

Follow-up of cases was conducted using a mixed method, through linkage with the National Index of Mortality (Indice Nacional de Defunciones) and through active case finding (i.e., revision of clinical records) of cancer cases with poor prognosis and advanced age. Time of follow-up was defined as the time interval elapsed between the date of diagnosis and the date of death (for deceased patients), the date of last contact recorded (for patients lost at follow-up) or 31 December, 2021 (for the rest). Mortality data were retrieved from the information system of the Spanish Ministry of Health, considering the same ICD-O-3 codes and study period employed for incidence data.

To study the incidence of anogenital cancers, the following variables were included: age at diagnosis, sex, anatomical site, tumor morphology and behavior according to ICD-O-3, most valid data source for diagnosis, date of diagnosis, date of last contact recorded, and vital status. Regarding mortality, age group, sex, ICD-10 anatomical site and year of death were included.

Age at diagnosis was categorized in 10-year intervals for the specific analysis: <25, 25–34, 35–44, 44–54, 55–64 and ≥65 years. The study period was divided into three time periods: from 1985 to 1997, from 1998 to 2007, and from 2008 to 2017, as we hypothesized that sexual behaviors—which influence the risk of HPV infection –, may have varied over an extended period of time. Additionally, this enabled the estimation of epidemiological indicators for more homogeneous cases of anogenital cancer (hence facilitating inter-period comparisons).

The crude and age-standardized incidence rates were estimated by the direct method using the standard European (ASIR-E) and World (ASIR-W) populations. In the analysis as a function of age group, the specific incidence rates were estimated.

In the time-trend analysis, the log-linear joinpoint regression method was applied to the age-standardized incidence rates (ASIR-E) in each year of incidence (26). This technique estimates the annual percent change (APC) and its 95% confidence interval, along with the time points where the trend significantly changes.

Analogous to the incidence study, crude (CMR) and adjusted (ASMR-E and ASMR-W) rates were calculated for mortality data, and a time-trend analysis was also conducted.

We calculated overall survival, which considers all causes of death, and net survival, defined as the survival that would be observed if cancer were the only possible cause of death. Overall survival with 95% confidence intervals was obtained by the Kaplan-Meier method; net survival was obtained using the Pohar-Perme estimator (27), with life tables for competing death probabilities (according to year, sex and age), smoothed by the Elandt-Johnson method (28). In order to draw comparisons between time periods and between sexes, net survival was further standardized by age, utilizing the standard population of patients with cancer (28).

All statistical analyses were performed using Stata software version 17 (StataCorp. 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC.) with the exception of trend analysis, for which Joinpoint software version 4.9 was used (Joinpoint Regression Program, Version 4.9.1.0–April 2022; Statistical Methodology and Applications Branch, Surveillance Research Program, National Cancer Institute).

Between 1985 and 2017, a total of 1,951 cases of cancer of the anogenital area were registered. These included 1,112 cases of cervical cancer, 205 cases of anal cancer, 234 cases of penile cancer, 343 cases of vulvar cancer, and 57 cases of vaginal cancer (see Table 1). The trend analysis of the standardized incidence ratio (ASIR-E) showed no statistically significant changes for any cancer site during the study period (see Figure 1). The APC was overall negative, with some variation according to anatomical site. For cervical cancer, there was a slightly decreasing trend (APC = −0.45%; 95% CI: −1.2 to 0.3), similar to that of anal cancer (APC = −1.03%; 95% CI: −2.8 to 0.8), penile cancer (APC = −2.07%; 95% CI: −4.3 to 0.2), and vaginal cancer (APC = −1.49%; 95% CI: −9.2 to 6.9); in contrast, an increasing trend was observed for vulvar cancer (APC = 1.03%; 95% CI: −0.4 to 2.5). Were these trends to continue, the estimated ASIR-E of anogenital cancers by 2022 would be 1.7 per 100,000 men and 8.0 per 100,000 women (Supplementary Figure S1).

Overall, four out of five incident cancers of the anogenital area (82.1%) occurred in females, as cervical cancer was the most frequent malignancy, responsible for 57.0% of all cases. The second most frequent cancer was vulvar cancer, and the least frequent was vaginal cancer (2.9%). In men, penile cancer was the most frequent with 234 cases (12.0%), followed by anal cancer (5.4%).

Higher incidence was found for individuals ≥65 years old for all cancer sites (irrespective of sex). The one exception was cervical cancer, where the highest incidence occurred in women 45–64 years old (see Figure 2). More than 90% of vulvar cancers were diagnosed in women aged 55 years and above, and most vaginal cancers also occurred in women ≥65 years old. Likewise, two thirds of cases of penile cancer occurred in men ≥65 years old. With regard to anal cancer, men were more frequently affected (53.4%) compared to women, particularly those aged 55 years and over; for both sexes, incidence increased with age. In terms of the spatial distribution, the ASIR-E of anogenital cancers between 2008 and 2017 ranged from 9.3 to 10.5 per 100,000 women and from 1.9 to 2.5 per 100,000 men across health districts (Supplementary Figure S2). Of note, the Granada district (which includes the province's capital city) showed the highest ASIR-E in women and the second highest ASIR-E in men; at any rate, the variability observed was quite low.

During the study period, there were 532 fatalities caused by the anogenital cancers studied: 311 were due to cervical cancer (58.5%), followed by vulvar (22.4%), and penile cancer (8.1%) (see Table 2). Similar to the sex distribution of incident cancers, women accounted for 88% of the deaths registered.

Figure 3 shows the trend analysis of the standardized mortality rates (ASMR-E), which yielded a statistically significant decreasing trend for cervical cancer, going down from 2.26 in the first time interval (1985–1997) to 1.43 in the third (2008–2017). This resulted in a statistically significant APC of −3.48% (95% CI: −5.1 to −1.8) from 1991 to 2017 (an accused growing trend was observed prior to 1991). For the remaining cancer sites, the APC trend in the study period was increasing, albeit not significantly.

Mortality globally increased with more advanced age for all cancer sites (see Figure 4). Women of advanced age, particularly above 65 years, had higher mortality for cervical, vaginal, and vulvar cancer (Figure 4). In women, anal cancer mortality was higher among those ≥65 years (specific mortality rate = 0.46); in contrast, in men, age differences in anal cancer mortality were much smaller.

As illustrated in Figure 5, cervical and vulvar cancer showed an upward trend in 5-year survival between 1985 and 1997 and 2008–2017, with a 9.9% increase (from 57.4% to 67.3%) and a 13.8% increase (from 54.1% to 67.3%), respectively. For penile cancer an irregular evolution was observed, with an estimated 5-year survival of 68% in the second period compared to 100% in the first period, and then increasing to 74% in the last period. However, no statistically significant changes were observed for any cancer site.

The 1-year, 3-year, and 5-year survival rates (including overall survival, net survival, and age-standardized net survival) in the period 2008–2017 for all cancer sites studied are displayed in Table 3. Anal cancer survival rates stratified by sex are further detailed in Supplementary Table S1.

In the current study, half of all cervical cancers affected women aged 45–64 years, consistent with data from other population-based cancer registries (31, 32). An overall decline was observed in cervical cancer incidence since 1985, but not as marked as the trends previously described in Spain (33, 34), the United States (35), China (36), and Poland (37). This discrepancy may stem from the low regional incidence of cervical cancer at the start of the study in comparison to other geographical locations (32, 37). Another potential factor is compliance with the cervical cancer screening program, formally included in the Spanish National Health System in 2014 (38); Andalusia, the autonomous community to which Granada belongs, is among those with poorer compliance (39). In fact, according to the incidence rates obtained between 2008 and 2017, Granada would currently present an incidence above the average in Spain (25).

The increasing trend in mortality observed during the first six years of the study period could be due to the absence of cervical cancer screening programs and the lower accuracy of imaging tests, which led to diagnosis in more advanced stages; and by the limited effectiveness of the treatment available at the time. Since then, Bosetti et al. (40) identified the decline in cervical cancer mortality as one of the major contributors in the reduction of cancer mortality in Europe. We found a decreasing trend for cervical cancer mortality, reaching an ASMR-E = 1.43 deaths per 100,000 women between 2008 and 2017, similar to the values reported in Northern and Western Europe (11, 32, 41). This finding may be explained by cervical cancer screening, which has shown to decrease cervical cancer mortality in Europe (42), and by the inclusion of chemotherapy in the therapeutic arsenal for advanced cervical cancer (43). Moreover, immunotherapy might further reduce cervical cancer mortality in upcoming years (44). The 5-year age-standardized net survival (taking the European population as reference) was estimated at 67%, in accordance with the rates reported in other studies (30, 33, 45).

Vulvar cancer represented between 15% and 20% (in 1985–1997 and 2008–2017, respectively) of all anogenital cancers. A smaller contribution, usually around 10%, has been found in similar studies (29, 46, 47). The incidence of vulvar cancer found in our study is comparable to that found in other high-income countries, although slightly higher than the incidence reported in studies from Denmark and South Korea (46, 48) and slightly lower than in studies from Germany and Japan (49, 50). Most vulvar cancers were diagnosed in women ≥65 years old, consistent with findings from other populations (51, 52).

Vulvar cancer was also ranked second in terms of mortality and the ASMR-E in our population was lower than the one found in other European countries (53). It is possible that prognostic factors as the tumor stage at diagnosis (not included in our study) underlie these differences. There was little variation in survival throughout the period analyzed, with a 5-year ASNS of 64%, concordant with global estimates (52) and with the period-specific survival rates recently described in the United States (54).

The anogenital cancer with the lowest incidence and mortality was vaginal cancer; yet, its mortality increased, especially during the last decade analyzed. The incidence and mortality rates observed are very similar to those found in Denmark (46). However, both higher incidence (30) and higher mortality (as much as 3-fold, with an ASMR-W of 0.3 per 100,000 women) (48) have been reported for other countries. Almost all cases occurred in women older than 65 and this was the age group with highest mortality, similar to previous studies (22). The survival rates for vaginal cancer in our population are below those reported in other countries (54). Stage at diagnosis, which was not included in the survival analysis, may account for these differences.

The epidemiological indicators for anal cancer showed some differences with previous research. Anal cancer ranked fourth in terms of incidence, in contrast to studies from other high-income countries that have found it is the second most frequent anogenital cancer (46, 47, 55). Studies from several other countries have reported higher age-standardized incidence rates than those found in the current study (30, 31) and rising incidence trends (56), something that was not observed in our study. We found the highest incidence in the oldest age groups, whereas several studies from other countries have reported highest incidence for individuals aged 45 to 69 (15, 57). In our study, anal cancer was more frequent in men than in women, as described in previous studies analyzing nationwide data (17, 25).

Almost two thirds of deaths from anal cancer occurred in men and mortality increased with age. Mortality followed an upward trend, but was still well below rates at the national (17) and global level (58). With respect to survival, the 5-year survival rates observed in this study were lower than in comparable populations (54, 59). Survival rates, which are influenced by stage at diagnosis and type of treatment, among others, may vary between 80% in localized disease and 20% in metastatic cancer.

We found higher incidence rates for penile cancer compared to North America (47, 55). We also observed a slight decrease in incidence, in contrast to other studies showing stable or increasing temporal trends (19, 56, 60). However, the incidence rates obtained (ASIR-E and ASIR-W) were similar to those found in cancer registries in other European countries, that have a similar prevalence of circumcision (25, 61). We also found that penile cancer mortality increased during the study period, with rates concordant with data from other countries (19, 61). Penis was the cancer site with second highest survival between 2008 and 2017: its 5-year survival rate exceeded 70%, similar to values reported for countries from Northern Europe (62).

HPV infection is one of the most frequent sexually-transmitted infection worldwide (63). The prevalence of genital HPV infection shows great variation across populations, ranging from 2% to 45%, and is higher in less-developed regions, namely in Oceania and Africa (64). Such large differences are related to several risk and protective factors that we now briefly summarize.

Subjects with a weakened immune system are more susceptible to HPV infection. In particular, those infected by human immunodeficiency virus (HIV) or undergoing immunosuppressive therapies (for instance, in the context of solid-organ transplantation) are at higher risk (65). Conversely, vaccination against HPV, especially with the 9-valent vaccine, is highly effective at reducing the incidence of HPV infection (66). Besides, tobacco smoking has also shown to increase the risk of acquiring HPV (67). Furthermore, psychosocial aspects may be even more relevant. Since HPV is primarily transmitted through sexual contact, factors such as early age at sexual debut, high number of sexual partners, risky sexual behaviors, and measures to improve risk perception (for instance, by providing adequate sexual education to young and adolescent individuals) or lack thereof all influence the risk of exposure to HPV and of a subsequent potential infection (68).

Many of the abovementioned factors show an uneven global distribution. HIV infection still disproportionately affects sub-Saharan Africa; about 15% of girls in the target age for HPV vaccination are fully immunized (69); the overall burden of tobacco consumption is mainly carried by low- and middle-income countries (70); multiple countries have room for improvement in terms of education (71), let alone sexual education.

To the best of our knowledge, this is the first study to simultaneously analyze incidence, mortality and survival trends for all anogenital cancer sites over a period longer than 30 years. However, it presents some limitations. Firstly, with the exception of cervical cancer, the cancers studied are relatively rare. In addition, the population from which the Granada Cancer Registry draws data is also relatively small (930,000 inhabitants). As a result, the number of cases is generally insufficient to reach high statistical power, increasing the probability of type 2 errors (failing to find significant differences when these actually exist). Secondly, multiple factors involved in the etiology of anogenital cancer, such as the prevalence of HPV infection, a previous history of anogenital cancer, and sociocultural determinants, were not included. Although vaccination against HPV may influence epidemiological data, we believe that it has not affected our results, because it was not included in the regional systematic vaccination schedule until 2008 (when it was also restricted to females aged 14 years) and the period elapsed between HPV infection and anogenital cancer development can range from years to several decades. Thirdly, while this study covers five different cancer types, the epidemiology of HPV infection is not limited to the anogenital area—HPV has also been involved in the development of other cancers, namely oropharyngeal cancer (72). Consequently, the total contribution of this pathogen to the cancer burden is to some extent greater than presented here. However, the decision to restrict the cancer sites analyzed to those in the anogenital area increased the comparability of our estimates. Finally, we were unable to adjust for key prognostic factors like the tumor stage at diagnosis or the type of treatment.