The definition of infertility is the failure to conceive after 1 year or more of regular intercourse without taking contraceptive measures (Practice Committee of the American Society for Reproductive Medicine, 2008). Cancer treatment-related infertility may be transient, permanent or delayed [i.e. premature ovarian failure (POF) in women]. The main mechanisms are primary hypogonadism as a result of testicular or ovarian damage after cytotoxic treatment or radiotherapy, secondary hypogonadism due to pituitary or hypothalamic dysfunction after cranial irradiation, and structural damage to the uterus caused by pelvic irradiation. Fertility considerations differ between males and females due to different mechanisms of gonadal dysfunction and also because of the different nature of fertility preservation techniques. In women, cancer treatment causes POF, which not only results in infertility but also endocrine disorders, such as oestrogen deficiency. In men, on the other hand, because the germinal epithelium of the testes is more susceptible to cytotoxic damage than the Leydig cell, infertility is a relatively frequent consequence, but testosterone deficiency is less common (Bramswig et al, 1990; Sieniawski et al, 2008a). Each haematological malignancy has a unique constellation of fertility considerations that relates to the disease itself, the gonadotoxic potential of common treatment protocols, and the age of the patient population.
In this section we will discuss data relevant to both NHL and HL and will underline the main differences in fertility considerations between these two diseases. The specific treatment modality of HSCT is discussed elsewhere in this review.
Adult males. General issues: Semen analysis is the most reliable non-invasive test of fertility in males (Barrat, 2007). Elevated follicular stimulating hormone (FSH) levels are often employed as indirect markers of testicular dysfunction but their reliability is questionable and conflicting reports exist (Sy Ortin et al, 1990; Rueffer et al, 2001; Sieniawski et al, 2008a,b). Inhibin B, which is produced by Sertoli cells, is another indirect marker of male fecundity (Mabeck et al, 2005) and decreased Inhibin B levels have also been associated with impaired spermatogenesis in children and adults receiving chemotherapy (van Beek et al, 2007). These indirect markers are especially useful in evaluating gonadal dysfunction among pre-pubertal males.
The pre-treatment semen quality of patients with HL and NHL has been shown to be significantly inferior when compared to healthy controls (Shekkariz et al, 1995; Hallak et al, 2000). In one representative study of pre-treatment semen quality in HL patients, the spectrum of dyspermia ranged from severe damage such as azoospermia in 21% to moderate and minor disorders in a further 49% of subjects (Rueffer et al, 2001). In the same study and many others (Viviani et al, 1991; Sieniawski et al, 2008a) only up to 30% of newly diagnosed HL patients had sperm quality within the normal range, albeit another report exhibited normospermia in as many as 66% (Blackhall et al, 2002).
The mechanism of this dysfunction is unknown and may be due to incidental factors or systemic symptoms like fever (Marmor et al, 1986). It has been hypothesized that the lymphoma itself has a direct effect on spermatogenesis, with several pro-inflammatory cytokines [e.g. Interleukin 1 (IL1) 1, IL6, tumour necrosis factor-α, and soluble interleukin receptors 2 and 6 (sIL2R and sIL6R respectively)] proposed as possible mediators of infertility (Rueffer et al, 2001). Although cytokines and inflammation have been implicated in male infertility in a general setting (Hales et al, 1999; Fraczek & Kurpisz, 2007), their role specifically in lymphoma-induced infertility is purely hypothetical on the grounds that lymphoma is a biologically active disease. This hypothesis is somewhat supported by Rueffer et al (2001) who indentified elevated erythrocyte sedimentation rate and advanced stage of disease as risk factors for sperm damage among pre-treatment HL patients. There are however reports challenging this claim, showing no association between the disease stage and quality of semen in HL (Viviani et al, 1991; Shekkariz et al, 1995; Hallak et al, 2000; Sieniawski et al, 2008a). The majority of data available on the effect of chemotherapy on gonadal function in lymphoma is derived from patients receiving gonadatoxic treatment with alkylating agents. Although gonadal function deteriorates in the majority of patients, a small minority experience improved sperm counts (Sieniawski et al, 2008a). Furthermore, the fertility status at diagnosis does not appear to predict post-treatment fertility, although this presumption is based on limited data, all from patients treated with alkylating agents (Viviani et al, 1991; Sieniawski et al, 2008a,b). These findings raise the question whether pre-treatment dyspermia in lymphoma is reversible following treatment with regimens that hold low gonadotoxic potential. There are no comprehensive data available on a direct comparison between sperm counts prior to and after treatment with such regimens (e.g. ABVD) in lymphoma and subsequently this interesting subject warrants further research.
Specific effects of different treatment regimens: The majority of the literature focuses on treatment-induced infertility in HL, with a paucity of trials focusing solely on NHL. Statistics pertaining to the effect of specific treatment protocols on fertility are summarized in Table I. There are several factors that influence the incidence and degree of testicular dysfunction in patients receiving definitive treatment for lymphoma. The type of treatment used has an influence on the degree of gonadal dysfunction (Kiserud et al, 2009). Chemotherapy-induced gonadal failure is most often caused by alkylating agents, among which procarbazine and cyclophosphamide are considered the main culprits. The vast majority of adult males receiving therapy containing alkylating agents for HL have long-lasting or permanent azoospermia (Viviani et al, 1991; Blackhall et al, 2002; Sieniawski et al, 2008a). Recently, equally high rates of azoospermia of 89% were shown in 38 patients with advanced-stage HL treated with baseline or escalated BEACOPP, a treatment regimen comprised of bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine and prednisone (Sieniawski et al, 2008b). In contrast, the ABVD regimen, which contains no alkylating agents, causes only temporary azoospermia that is typically reversible in all patients (Viviani et al, 1985; Anselmo et al, 1990; Sieniawski et al, 2008a). van der Kaaij et al (2007) demonstrated a similarly negligible rate of gonadal dysfunction (albeit estimated with the surrogate marker, FSH) with ABVD or EBVP (epirubicin, bleomycin, vinblastine and prednisone) treatment for early-stage HL, measured as 8% [95% confidence interval (CI): 3–15%, n = 101] after a median of 35 months. This rate was significantly lower than rates observed in patients treated with regimes containing alkylating agents, such as BEACOPP. Other treatment protocols devoid of alkylating agents also allow substantial post-treatment recovery of spermatogenesis (Meistrich et al, 1997).
Table I. Risk of infertility per treatment protocol, among males receiving treatment for haematological malignancies.
|VACOP-B or MACOP-B||NHL||Adults||15|| || ||0% (28 months)||Retrospective||(Muller & Stahel, 1993)|
|MOPP, pelvic radiation or both||HL||Children||12||NA||17% (9 years)||83% (9 years)||Retrospective||(Sy Ortin et al, 1990)|
|MOPP ± radiation||HL||Children||19||NA||0% (14 years)||95% (10 years)||Retrospective||(Heikens et al, 1996)|
|MVPP or ChlVPP/EVA ± radiation||HL||Adults||74|| || ||97% (33 months)||Retrospective||(Blackhall et al, 2002)|
|Various chemotherapy regimens ± radiotherapy||HL||Adults||269||45%||63% of those attempting (NA)|| ||Retrospective||(Kiserud et al, 2007)|
|Radiotherapy alone||HL||Adults||9|| || ||11% (17·4 months)||Prospective||(Sieniawski et al, 2008a)|
|BEACOPP or COPP/ABVD alone||HL||Adults||10|| || ||90% (17·4 months)||Prospective||(Sieniawski et al, 2008a)|
|BEACOPP or COPP/ABVD, with irradiation||HL||Adults||93|| || ||67% (17·4 months)||Prospective||(Sieniawski et al, 2008a)|
|BEACOPP baseline or escalated||HL||Adults||38|| || ||89% (NA)||Prospective||(Sieniawski et al, 2008b)|
|Chemotherapy including cyclophosphamide||ALL||Children||11|| || ||73% (13 years)||Retrospective||(Marquis et al, 2010)|
|Cyclophosphamide-based conditioning for HSCT||SAA, ALL, AML, CML, lymphoma||Adults||618||NA||5·3% (NA)|| ||Retrospective||(Sanders et al, 1996)|
|Cyclophosphamide-based conditioning for HSCT||NHL, HL, AML, ALL, SAA||Adults||64|| || ||70% (NA)||Retrospective||(Anserini et al, 2002)|
|Autologous HSCT||NHL, HL, AML, ALL, SAA, CML||Adults||119||NA||7% (7·2 years)|| ||Retrospective||(Carter et al, 2006)|
|Allogeneic HSCT|| ||Adults||208||NA||9% (13·6 years)|| ||Retrospective||(Carter et al, 2006)|
Higher cumulative doses of cytotoxic drugs, especially alkylating agents, also correlate with a higher proportion of permanent azoospermia among males receiving treatment for HL (da Cunha et al, 1984; van der Kaaij et al, 2007; Sieniawski et al, 2008a) and NHL (Pryzant et al, 1993). This conclusion was reached by comparing the degree of fertility impairment between various chemotherapy and radiotherapy treatment groups. However, in most of these studies, these subsets of low, medium and high gonadotoxicity were created in a qualitative or semi-quantitive manner and exact doses of chemotherapeutic agents were not published. For example, Kiserud et al (2007) retrospectively evaluated the probability of post-treatment parenthood in 120 males attempting parenthood after treatment for HL. In this study chemotherapy regimens with low gonadotoxicity were devoid of alkylating agents; protocols with four or less courses containing alkylating agents were classified as having medium gonadotoxicity; and more than four courses with alkylating agents or conditioning chemotherapy before HSCT were considered highly gonadotoxic. The 10-year probability of achieving post-treatment parenthood was 85%, 35% and 18%, respectively. Another analysis of HL patients after treatment using elevated FSH as a surrogate marker, demonstrated a possible effect on fertility after three cycles of BEACOPP (or other regimens with equivalent doses of alkylating agents). FSH was higher in this group than in subjects who received no alkylating agents and this effect increased with each dose increment. Moreover, higher doses of alkylating agents were associated with a lower rate of FSH normalisation (van der Kaaij et al, 2007). Other studies have used similar models (da Cunha et al, 1984; Sieniawski et al, 2008a) to show a graded effect of cumulative doses of alkylating agents on fertility. Even so, there is no clear cumulative dose of alkylating agents above which gonadal damage is caused. However, the gonadotoxic potential of these agents appears to reach a plateau above a certain cumulative dose, which may explain why no difference in azoospermia rates was found between patients receiving eight cycles of BEACOPP baseline and those given eight cycles of BEACOPP escalated (Sieniawski et al, 2008b). These treatment regimens contain the same quantity of procarbazine (5600 mg/m2) but the latter incorporates twice the cumulative dose of cyclophosphamide than the former (10 000 vs. 5200 mg/m2).
Another factor associated with impaired spermatogenesis in patients with HL and NHL is pelvic and abdominal irradiation. (da Cunha et al, 1984; Pryzant et al, 1993). The degree of impairment and duration thereof is associated with testicular radiation doses and the precise infra-diaphragmatic radiation fields employed, with more inferior fields carrying a higher risk (Dubey et al, 2000). Specifically, cumulative gonadal doses of only 15 cGy significantly reduce the sperm count. However, azoospermia is usually not observed with doses below 70 cGy, whereas gonadal doses of 200–300 cGy cause prolonged and often irreversible azoospermia. Moreover, pelvic radiotherapy has a further deleterious effect on gonadal function when combined with chemotherapy containing alkylating agents (da Cunha et al, 1984; Pryzant et al, 1993).
Notably among adult male HL patients, post-treatment fertility is not related to the age at treatment (Sieniawski et al, 2008a), in contrast with the age-related associations seen in female HL patients. However in one study, a sub-group analysis of male HL patients over the age of 50 years demonstrated significantly higher post-treatment FSH levels than in younger counterparts (van der Kaaij et al, 2007).
Patients rendered azoospermic during HL and NHL treatment may potentially regain spermatogenic capacity at varying times after treatment, even over a period of several years. Furthermore, the timing of the onset of recovery does not necessarily correlate with the period of time needed to regain full fertility (Pryzant et al, 1993). Higher doses of alkylating agents are associated with a prolonged interval before resumption of gonadal function (van der Kaaij et al, 2007). The rate of recovery is elucidated by prospective data on male fertility in patients with advanced-stage HL receiving BEACOPP and COPP (cyclophosphamide, vincristine, procarbazine and prednisone)/ABVD-based regimes, which showed that the median time to recovery of spermatogenesis was 27 months (Sieniawski et al, 2008a). Among those patients who regained spermatogenesis, the incidence of gonadal recovery increased every year, from an incidence of 18% during the first post-treatment year to 35% after the third year. A shorter median time to recovery, albeit based on FSH levels alone and not on sperm counts, has been shown among early-stage HL patients treated with ABVD in comparison with those treated with alkylating agents (van der Kaaij et al, 2007).
Finally, because sperm count is considered a reliable predictor of the ability to father a child, it is a commonly used end-point for gauging fertility after cancer treatment, as demonstrated in the aforementioned data. There is however a lack of large trials evaluating the important end-point of post-treatment fatherhood. In one of the largest studies evaluating post-treatment parenthood in HL survivors, 45% of the 269 men enrolled attempted parenthood. Sixty-three percent of these men achieved fatherhood without the use of assisted reproductive techniques (ART). ART resulted in parenthood in 10 additional cases (Kiserud et al, 2007).
Importantly, the outcome of post-treatment parenthood has several shortcomings that apply to both sexes, irrelevant of the disease treated: It is affected by the percentage of patients attempting parenthood, which is not always reported (see Table I); this sequentially depends upon interconnected factors such as age, marital status, the number of children prior to treatment, the patient’s perception of fertility potential and the desire to become a parent; the patient’s partner may also have confounding fertility issues; most studies do not relate to the time required to achieve parenthood. Therefore, this outcome should be interpreted with caution, but it can be concluded that it may or may not be possible to have a live born infant following chemotherapy.
Adult females. General issues: In females, cancer treatment-induced gonadal failure may manifest itself as acute ovarian failure or chemotherapy-associated amenorrhea soon after completion of the treatment regimen. A proportion of these patients resumes normal menstruation and regains fertility within a timeframe of months to years (Clark et al, 1995). Alternatively, the damage caused may only become apparent years later as POF, which is defined as amenorrhea for more than 6 months with elevated FSH levels, in women under the age of 40 years. Women have a limited oocyte reserve that decreases with age until menopause, and is directly damaged by cytotoxic treatment via apoptotic cell death and ovarian atrophy with loss of primordial follicles (Familiari et al, 1993). Therefore the risk of chemotherapy-induced POF increases with age at first treatment, which has been shown to be an independent risk factor for POF in HL (De Bruin et al, 2008). Moreover, some studies of female HL and NHL patients specify a significant increase in risk above the age of 30 years, but this was an arbitrary cut-off age (Franchi-Rezgui et al, 2003; Fossa & Magelssen, 2004; Behringer et al, 2005; Kiserud et al, 2007). This repeated finding of age-related gonadal injury after treatment in women may also be influenced by a natural decline in fertility with increased age. In contrast with male HL patients, there is no clear evidence of pre-treatment fertility impairment. For example, in one study 89·6% of women had regular pre-treatment menstrual cycles. (Behringer et al, 2005).
Specific effects of different treatment regimens: In congruence with their effect in males, treatment with alkylating agents especially procarbazine and cyclophosphamide, and higher cumulative doses thereof, are associated with an increased risk of POF in women receiving treatment for HL and NHL (Franchi-Rezgui et al, 2003; De Bruin et al, 2008). One research group calculated an 11% increase in risk of POF with each additional 1·4 g/m2 of cumulative procarbazine administered as part of treatment for HL. After multivariate logistic regression, cyclophosphamide [Hazard Ratio (HR): 3·5; 95% CI: 2·0–5·9] and procarbazine (HR: 8·1; 95% CI: 2·0–32·8) but not mechlorethamine, carried an increased risk of POF when compared with patients who received radiotherapy alone, which didn’t include the ovaries in the radiation field (age range at treatment: 14–39 years). This association was not seen in a sub-group of patients who received <4·2 g/m2 of cumulative procarbazine (De Bruin et al, 2008). This is the first study to evaluate the dose dependent risks for specific agents separately but others have assessed this risk per-protocol. For instance, 51·4% of female HL survivors (n = 74, age: 15–40 years) who received eight cycles of escalated BEACOPP had continuous amenorrhea at median follow up of 3·2 years. This dosing regimen caused a higher rate of amenorrhea than standard BEACOPP dosing (Behringer et al, 2005), in contrast with the effect in males where no additional effect on fertility was shown (Sieniawski et al, 2008b). Other studies have shown a similarly high prevalence of premature menopause, around 50%, after treatment with alkylating agents. (Blumenfeld et al, 2008; De Bruin et al, 2008). Treatment protocols like ABVD and EBVP for HL, devoid of any alkylating agents, pose no documented risk of POF (De Bruin et al, 2008) and seem to result in normal rates of pregnancy (Hodgson et al, 2007). Although the majority of the above data pertains to HL there are several reports describing protocol-specific risks of infertility in NHL (see Table II). Females under the age of 40 years receiving CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) or CHOP-based chemotherapy for NHL have a very low risk of developing gonadal dysfunction (Muller & Stahel, 1993; Elis et al, 2006). There is also limited data suggesting that an intensified regimen of CHOP (Mega-CHOP) in NHL may cause similarly low rates of gonadal dysfunction (Dann et al, 2005). In light of the above data, despite causing a significant risk of infertility in women, alkylating agents seem to have a less devastating effect than in men.
Table II. Risk of infertility per treatment protocol, among females receiving treatment for haematological malignancies.
|VACOP-B or MACOP-B||NHL||Adults||7|| || ||14% (28 months)||Retrospective||(Muller & Stahel, 1993)|
|CHOP||NHL||Adults||36||NA||50% (84 months)||5% (84 months)||Retrospective||(Elis et al, 2006)|
|Hyper-CVAD||NHL, ALL||Adults||7||NA||43% (27 months)||14% (27 months)||Retrospective||(Seshadri et al, 2006)|
|MOPP, pelvic radiation or both||HL||Children||86||NA||28% (9 years)||19% (9 years)||Retrospective||(Sy Ortin et al, 1990)|
|ABVD||HL||Adults||36||100%||70% (12 months)|| ||Retrospective||(Hodgson et al, 2007)|
|Various chemotherapy regimens ± radiotherapy||HL||Adults||184||50%||75% (NA)|| ||Retrospective||(Kiserud et al, 2007)|
|Cyclophosphamide-based conditioning for HSCT||SAA, ALL, AML, CML, lymphoma||Adults||708||NA||4·5% (NA)|| ||Retrospective||(Sanders et al, 1996)|
|Autologous HSCT||NHL, HL, AML, ALL, SAA, CML||Adults||122||NA||3% of those attempting (11·2 years)|| ||Retrospective||(Carter et al, 2006)|
|Allogeneic HSCT|| ||Adults||170||NA||2% of those attempting (10·2 years)|| ||Retrospective||(Carter et al, 2006)|
Abdominal or pelvic radiation also raises the probability of POF. This risk is significantly raised when the ovaries receive direct radiation (De Bruin et al, 2008). There is a reverse correlation between age at treatment and the dose of radiation needed to cause POF. Wallace et al (2005) designed a model that predicts the expected age of ovarian failure after treatment with a known dose of radiation. Ovarian radiation doses below 400 cGy usually do not result in permanent ovarian dysfunction (Sy Ortin et al, 1990), compared with a lower threshold for significant gonadal damage in men (Dubey et al, 2000), which underlines the difference in susceptibility to radiation injury between the testis and ovary (Byrne, 1999).
It is important to note that the majority of the data available is retrospective and that most researchers use menopause after varying follow up periods, and not necessarily POF, as the main marker of fertility. De Bruin et al (2008) shed light on the risk of developing post-treatment POF in one of the largest retrospective cohorts to date of female HL survivors (age range at treatment: 14–39 years). POF was documented in 19% of the 518 women at a mean age of 33·5 years (range: 19–39 years) after a median follow up period of 9·4 years after treatment. (De Bruin et al, 2008). Importantly, only 3% of those who developed POF reached menopause within the first year after treatment, demonstrating the delayed clinical effect of HL treatment on the female reproductive system. This supports the fact that a regular menstrual cycle after cancer treatment is only a surrogate marker of parenthood potential and may still be followed by POF. Therefore pregnancy rates after lymphoma treatment are of great interest, though there is a shortage of reports evaluating this outcome. In one of the largest studies evaluating post-treatment parenthood in HL survivors, 50% of the 184 women enrolled attempted pregnancy (n = 91), and 75% of these women (median age at diagnosis: 23 years; range: 9·1–37·1 years) successfully gave birth without the use of ART. ART resulted in only one additional pregnancy (Kiserud et al, 2007).
Special considerations in children. Male and female survivors of conventional chemotherapy-based treatment for childhood NHL are at low risk of infertility and impaired puberty (von der Weid, 2008). As in adults, there is limited data on fertility after treatment for NHL and the majority of data relates to HL.
Importantly, both sexes are not equally affected by HL treatment. Boys treated with combined modality treatment, including chemotherapy incorporating alkylating agents and pelvic radiation, are at a significantly higher risk of gonadal dysfunction than girls receiving the same treatment (Sy Ortin et al, 1990; Papadakis et al, 1999) similar to the gender-related differences seen in adults. Furthermore, childhood HL and NHL patients share other risk factors with adult patients, such as treatment with alkylating agents and pelvic radiation, and cumulative doses of these two modalities (Bramswig et al, 1990; Papadakis et al, 1999; van den Berg et al, 2004; Chemaitilly et al, 2006; von der Weid, 2008; Green et al, 2009). This is illustrated by a retrospective study of gonadal damage in 47 male and female subjects treated with six courses of MOPP, six courses of ABVD or three courses of MOPP\ABVD, without radiotherapy, for childhood HL. Patients treated with ABVD or MOPP\ABVD had minimal gonadal dysfunction, which was assessed by the surrogate marker FSH, while more than 80% of those treated with six MOPP courses had elevated FSH (van den Berg et al, 2004).
Male survivors of childhood HL therapy including those treated with alkylating agents, usually undergo normal pubertal development with normal secondary sexual characteristics, although sub-clinical damage to Leydig cell function [i.e. elevated luteinizing hormone (LH)] may occur (Bramswig et al, 1990; Papadakis et al, 1999). On the other hand, fertility may be impaired, especially when using regimens containing alkylating agents for HL in young boys, such as ChlVPP (chlorambucil, vinblastine, procarbazine and prednisolone), which can cause gonadal dysfunction in 80–90% of cases (Heikens et al, 1996; Mackie et al, 1996; van den Berg et al, 2004; von der Weid, 2008). However, gonadal dysfunction rarely seems to occur when alkylating agents are avoided in this population, as in ABVD (van Beek et al, 2007) and OEPA (vincristine, etoposide, prednisone and doxorubicin) regimens (Gerres et al, 1998). Preliminary data from small populations show that two cycles of OEPA treatment for boys with HL carry a negligible risk of gonadal dysfunction, as measured by elevated FSH (Gerres et al, 1998; Schellong et al, 1999) compared with a rate of 28·9% seen with two cycles of the OPPA regimen (vincristine, procarbazine, prednisone and doxorubicin) which contains the alkylating agent, procarbazine (Bramswig et al, 1990). When two cycles of OEPA treatment are followed by two or four cycles of COPP, the rate of gonadal dysfunction rises to 37·5% and 36·5%, respectively (Gerres et al, 1998). Younger age and hence a pre-pubertal state, does not seem to prevent gonadal failure in young males treated for childhood HL and NHL (Ben Arush et al, 2000; van Beek et al, 2007) and the rate of gonadal dysfunction is probably similar to that among adult males receiving lymphoma treatment.
Conversely, younger girls treated for HL with alkylating agents may only have increased FSH/LH levels in around 20% (Papadakis et al, 1999). However, treatment for childhood cancers stills carries a considerable risk of POF, even later in life (Byrne et al, 1992). Studies assessing the effect of childhood lymphoma treatment on females do not clearly relate to the pubertal state at the time of treatment. They do however show that younger girls are more likely to maintain normal menses than older girls (Sy Ortin et al, 1990; Chemaitilly et al, 2006), much like the age-related associations seen in adult women. For example, Chemaitilly et al (2006) assessed acute ovarian failure in 3390 survivors of childhood cancer (30·7% with HL and 8·8% with NHL). There was a higher rate of acute ovarian failure (n = 215) among patients who were older than 12 years at diagnosis (Odds Ratio: 1·8, 95% CI: 1·4–2·4) in comparison with those aged 12 years or less. Age at diagnosis remained an independent predictor of acute ovarian failure after multivariate analysis (Chemaitilly et al, 2006). Moreover, a sub-group analysis of a retrospective study of girls who had not experienced menarche at the time of treatment for HL, showed a similar age of subsequent menarche in comparison with that of the general population (Sy Ortin et al, 1990).
Regimens devoid of alkylating agents carry a far lower risk of gonadal dysfunction. This is illustrated by a prospective cohort of 110 children (35 girls) treated with VAMP (vinblastine, doxorubicin, methotrexate, prednisone) and involved field radiation for favourable-risk HL (median age at treatment: 13·3 years) which demonstrated normal menses in all girls of pubertal age at a median follow-up of 9·6 years (range: 1·7–15). Furthermore, 12 of those women successfully gave birth to healthy children (Donaldson et al, 2007).
Importantly, offspring of female survivors of childhood cancers who received pelvic irradiation have an increased risk of preterm birth, low birth weight or being small for gestational age (Green et al, 2009). It is imperative to limit the use of gonadotoxic agents in pre-pubertal patients, because of the lack of satisfactory techniques for preserving fertility in this age group. Notably, there is still a lack of comprehensive prospective data on fertility among adult survivors of childhood lymphoma undergoing contemporary treatment. Such trials are needed in the future and some are already underway (Korholz et al, 2007).
The rate of treatment-induced infertility in leukaemia patients depends upon whether HSCT, with its gonadotoxic conditioning regimens, is employed (Watson et al, 1999). HSCT is also used in the treatment of haematological malignancies other than leukaemia and most of the available trials assess its effect on heterogeneous populations, which include patients with lymphoma and aplastic anaemia, as well as leukaemia. Therefore we will elaborate on this topic separately, elsewhere in this review.
Adults. Male leukaemia patients have lower pre-treatment semen quality than healthy donors (Hallak et al, 1999). Nevertheless, this phenomenon has been less thoroughly researched than among lymphoma patients. Standard chemotherapy regimens used to treat ALL and AML have low gonadotoxic potential. Treatment with chemotherapy alone for ALL results in immediate impairment in spermatogenesis however the majority of men subsequently recover reproductive capacity (Kreuser et al, 1988). As in lymphoma patients, male endocrine gonadal function seems to be preserved. The reproductive function of women treated with chemotherapy alone for ALL seems to be mostly preserved (Kreuser et al, 1988). Furthermore, post-pubertal AML patients treated with chemotherapy alone infrequently have impaired fertility, compared with significantly higher rates among those who received allogeneic or autologous HSCT (Watson et al, 1999).
Special considerations in children. Survivors of childhood leukaemia treated with modern conventional therapy, without HSCT, are at a relatively low risk of infertility and delayed or impaired puberty. In ALL, contemporary treatment protocols entail lower doses of gonadotoxic agents, particularly cyclophosphamide, and thus are unlikely to cause infertility (Liesner et al, 1994; Byrne et al, 2004a,b; Brydoy et al, 2007; Nurmio et al, 2009). However, cyclophosphamide does carry a risk of gonadotoxicity, as shown by a recent study that evaluated fertility parameters in 11 male adult survivors of childhood ALL who received chemotherapy, which included a median cumulative cyclophosphamide dose of 4·091 g/m2 (median of 8 years at diagnosis, range: 2–17 years). The study subjects had low sperm counts in 73%, which were also significantly lower than counts of normal control subjects (Marquis et al, 2010). In AML, regimens devoid of alkylating agents are most commonly used and thus treatment-induced infertility may even be less common than in ALL (Waxman et al, 1983; Leung et al, 2000). Liesner et al (1994) presented interesting data from 25 subjects (15 females) who received chemotherapy for childhood AML (n = 23) and myelodysplastic syndrome. The median age at diagnosis was 3·5 years and after a median follow-up of 4 years, the uterine size was appropriate for age in all females and the ovarian volume was normal in the majority of the 10 patients who had this variable measured. (Liesner et al, 1994).
There are nevertheless several exceptions to the encouraging data above which are applicable mainly to survivors who were treated with outdated treatment protocols or required more aggressive treatment. High-dose cranial irradiation is associated with impaired fertility in boys who were treated for ALL before the age of 10 years (Byrne et al, 2004a). Cranial radiotherapy also causes fertility deficits in female survivors of childhood ALL, especially among those receiving the treatment around the time of menarche (Byrne et al, 2004b). In addition, direct irradiation of the testes due to leukaemic infiltration in ALL invariably causes both permanent azoospermia and Leydig cell dysfunction (Castillo et al, 1990).
Haematopoietic stem cell transplantation (HSCT)
This treatment modality is used in the treatment of both lymphoma and leukaemia. This section relates solely to gonadal dysfunction caused by HSCT. The fertility considerations relating to the diseases themselves and other therapeutic regimens used in their treatment, are reviewed above. It must be stressed that many of the studies on fertility after HSCT considered patients after allogeneic and autologous transplant as one group. Although the majority of studies analyse the contribution of each component of the conditioning regimen [e.g. total body irradiation (TBI)] towards gonadal dysfunction, they do not directly compare fertility between myeloablative and non-myeloablative conditioning regimens. Furthermore, the vast majority of patients included in these trials had already received chemotherapy prior to pre-transplant conditioning (Chatterjee et al, 1994).
Adult males. The rate of post-transplant infertility is greatly influenced by the gonadotoxic potential of the conditioning regimen. Myeloablative pre-transplant conditioning regimes are based on alkylating agents and/or TBI, both of which have been implicated in causing marked germ cell damage and infertility (Chatterjee et al, 1994). The degree of gonadal insult inflicted by both of these modalities is dose dependent (Apperley & Reddy, 1995). Furthermore, graft-versus-host disease (GVHD) has also been suggested as a mediator of impaired fertility (Grigg et al, 2000), but this has not yet been fully proven.
Azoospermia is found among in more than 70% of allogeneic HSCT survivors who underwent myeloablative conditioning, while the remainder usually experience varying degrees of impaired spermatogenesis (Anserini et al, 2002; Claessens et al, 2006). However, recovery of spermatogenesis may occur after as many as 9–10 years (Anserini et al, 2002). One study showed no correlation between age at treatment and recovery rate of spermatogenesis (Anserini et al, 2002), however another report found a lower probability of post-transplant fatherhood among men aged 30 years or more at the time of transplant (Carter et al, 2006).
TBI-based regimes cause more gonadal dysfunction (Anserini et al, 2002) and a lower rate of post-transplant parenthood (Carter et al, 2006) than chemotherapy alone. There is limited data suggesting that conditioning with cyclophosphamide (200 mg/kg) alone may maintain spermatogenesis in a large number of cases (Sanders et al, 1983; Anserini et al, 2002). However, there is no clear evidence that non-myeloablative conditioning protocols spare gonadal function, at least with short follow-up (Kyriacou et al, 2003). A number of clinical trials evaluating the efficacy and safety of non-myeloablative and reduced intensity conditioning are currently underway and will hopefully provide more prospective data on post-treatment gonadal function.
Finally, post-transplant fatherhood is uncommon and significantly reduced compared to the general population. Carter et al (2006) showed that among 327 male survivors of allogeneic and autologous HSCT, 8% sired children in comparison with 66% of sibling controls. Infants conceived spontaneously by survivors of HSCT have a similar proportion of congenital anomalies and comparable rate of stillbirths and miscarriages as the general population (Sanders et al, 1996; Salooja et al, 2001; Carter et al, 2006).
Adult females. As in males, the use and dosage of TBI (Carter et al, 2006) and alkylating agents are associated with the severity of gonadal dysfunction. Either or both of these highly gonadotoxic treatments are used in traditional myeloablative conditioning regimes which lead to POF in 70–100% of women receiving them (Tauchmanovàet al, 2003; Cheng et al, 2005). TBI is considered to be the most gonadotoxic component of conditioning regimens (Sanders et al, 1988) but the data available to date do not enable recommending one conditioning regimen over another for preserving fertility. However, promising data from a small retrospective survey demonstrated that HSCT survivors who received reduced intensity conditioning (RIC) had a significantly lower prevalence of POF (37·5%) than women who received conventional myeloablative conditioning regimens (Cheng et al, 2005). Subsequently, prospective trials are warranted in order to determine the exact role of RIC regimens on the overall incidence of premature ovarian failure. Type of transplant (allogeneic versus autologous) or previous treatment with alkylating agents have not yet been shown to affect the prevalence of POF (Tauchmanovàet al, 2003).
Although recovery of ovarian function is rare, it may occur years after transplant even if POF develops initially. The risk of gonadal damage after HSCT in women also increases with older age at the time of transplant (Tauchmanovàet al, 2003). Moreover, females have a lower rate of post-transplant parenthood than men (Carter et al, 2006). Due to a shortage in high-quality evidence, it is difficult to ascertain whether this is down to the difference in the nature and feasibility of ART between the two sexes.
Although rare, pregnancies in survivors of autologous or allogeneic HSCT are likely to have a successful outcome in over 80% of cases, and there is no evidence of increased congenital abnormalities (Salooja et al, 2001; Carter et al, 2006). However, pregnant survivors of allogeneic HSCT who received TBI conditioning have significantly higher rates of preterm deliveries, caesarean sections and low birth weight babies than their healthy counterparts (Sanders et al, 1996; Salooja et al, 2001). Another study however showed no such associations (Carter et al, 2006). There are conflicting reports on the risk of miscarriage and stillbirth in pregnant HSCT recipients (Sanders et al, 1996; Salooja et al, 2001; Carter et al, 2006). These complications are likely to be a feature of age at exposure to irradiation, with patients treated early in childhood, while the uterus is still small, being at highest risk. In large studies of mixed age populations these differences between studies may be lost. A proposed mechanism by which radiation may mediate such complications is by causing decreased uterine elasticity via intrauterine artery damage (Critchley et al, 1992). GVHD has also been implicated in causing uterine damage.
Finally, the annual birth rate is this population is considerably reduced in comparison with the general population (Salooja et al, 2001; Carter et al, 2006). This is highlighted by one retrospective cohort that revealed a 3% prevalence of motherhood in 292 allogeneic and autologous HSCT survivors compared to 72% among sibling controls (Carter et al, 2006).
Special considerations in children. The considerations in children are much the same as those noted above for adult patients. As in adult patients TBI has a deleterious effect on fertility, and a high proportion of female and male AML patients who received radiotherapy during childhood as part of their treatment protocol had subsequent impaired fertility (Leung et al, 2000). TBI and myeloablative chemotherapy before HSCT invariably causes gonadal dysfunction in males and females and may also preclude normal progression through puberty. Moreover, limited data suggests a negative effect on uterine growth and ovarian volume (Liesner et al, 1994). Unfortunately, there is a lack of large studies evaluating fertility in this specific patient population.
Other treatment protocols
Myeloproliferative disorders may also be seen in patients with reproductive potential. Some animal trials have demonstrated impaired spermatogenesis under imatinib treatment, especially during early life. This has led to the speculation that imatinib may cause reduced sperm counts in humans and one report exhibited oligospermia in a male treated with imatinib (Seshadri et al, 2004). Nevertheless, clinical experience has not yet shown this to be true (Hensley & Ford, 2003). In addition, imatinib inhibits kinases that are expressed in mammalian ovaries and are potentially important in oocyte and follicle growth and development (Hutt et al, 2006). A recent case report implicated imatinib in causing ovarian failure (Christopoulos et al, 2008). However, animal data has not showed impaired gonadal function in female animals and imatinib is not considered to impair fertility in female humans. To date there is insufficient data on the effect of second generation tyrosine kinase inhibitors on reproductive potential, although successful pregnancies after female and male use of these drugs have recently been reported (Cortes et al, 2008; Conchon et al, 2009; Oweini et al, 2011). It must be stressed that the aforementioned data is limited, based on small studies and some anecdotal reports, and therefore should be interpreted with caution.
In addition, there is extensive murine data, yet a paucity of case reports in humans, indicating that hydroxycarbamide may impair spermatogenesis (Garozzo et al, 2000; Grigg, 2007; Masood et al, 2007). Although there is significant clinical experience with the use of monoclonal antibodies, such as rituximab, their effect on gonadal function has not been documented and is thus not yet clear.