Fever in women may interfere with follicular development during controlled ovarian stimulation

Abstract

Purpose: This study aims to evaluate the effects of fever on follicular development in women undergoing controlled ovarian stimulation during in vitro fertilisation (IVF) and intracytoplasmic sperm injection (ICSI) cycles.

Materials and methods: This was a retrospective observational self-controlled study at a tertiary-care fertility centre. Six gonadotropin stimulation cycles characterised by poor ovarian response in which women reported the occurrence of a febrile illness, were considered for evaluation. Fever-exposed cycles were compared to the next stimulation cycle in the same women. Primary outcome measures were final number of pre-ovulatory follicles (≥ 16 mm) and final peak serum estradiol levels (pg/mL). Other outcome measures were final number of medium-sized follicles (12–15 mm), final mean estradiol serum level per follicle ≥ 12 mm (pg/mL), total days of stimulation and total gonadotropin ampoules utilised.

Results: Fever-exposed cycles were associated with significantly lower number of pre-ovulatory follicles (0.7 ± 0.8), significantly higher number of medium-size follicles (21.0 ± 4.5), and significantly reduced serum estradiol per follicle ≥12 mm (50.7 ± 11.7 pg/mL). They also required a significantly longer duration of ovarian stimulation (15.7 ± 3.3 days) and a significantly increased number of gonadotropin ampoules (47.2 ± 10.9). Four women had polycystic ovary syndrome and one hypothalamic hypogonadism.

Conclusion: This preliminary report suggests a possible negative effect of fever on follicular development and ovarian estradiol production in some women undergoing controlled ovarian stimulation.

• Hyperthermia
• febrile illness
• oocytes
• controlled ovarian stimulation
• follicular development

Introduction

Ovulatory dysfunction and embryonic mortality are known reproductive consequences of hyperthermia in animals. In fact, poor reproductive performance was observed by many investigators when animals were exposed to high ambient temperatures [1–4]. In vivo heat stress was demonstrated to impair ovarian steroidogenesis [5], [6], delay follicular development [5], [7], [8], accelerate regression of the corpus luteum [5], [7], [8], affect embryo development [9], [10], reduce implantation [11], and decrease embryo cleavage rates [9].

Similar effects have not been reported in humans and studies evaluating the effects of hyperthermia on follicular development in women are non-existent. In this comparative observational study, the effects of self-reported febrile illnesses on follicular development in women undergoing controlled ovarian stimulation during in vitro fertilisation (IVF) and intracytoplasmic sperm injection (ICSI) cycles were evaluated.

Materials and methods

Patients

A total of 325 consecutive gonadotropin-stimulated patient cycles for assisted reproductive technologies performed in an academic tertiary-care fertility centre were reviewed retrospectively over a period of 3 years. Cycles in which patients self-reported the occurrence of febrile illnesses were evaluated irrespective of women's age, duration of infertility or aetiology.

Institutional Review Board approval for the study was obtained.

Ovarian stimulation

Patients underwent pituitary desensitisation using the gonadotropin-releasing hormone (GnRH) agonist long suppression protocol (Decapeptyl, triptorelin, Ipsen Pharma, Dublin, Ireland) 0.05 mg subcutaneously daily starting the mid-luteal phase of the previous cycle. Women with hypothalamic hypogonadism did not receive GnRH agonist treatment. All patients were all stimulated using human menopausal gonadotropins (hMG) (Pergonal 75, Serono, Italy) or urinary follitropins (uFSH) (Metrodin 75, Serono, Italy) from day 3 of their menstrual cycle. The first ultrasound and estradiol controls were normally performed on cycle day 8, and gonadotropin dose was individually adjusted accordingly. Triggering of final oocyte maturation was decided when at least three leading follicles reached ≥18 mm in diameter. Human chorionic gonadotropins (hCG) (Pregnyl 5000, Organon, the Netherlands) were administered for final triggering as a single 10,000 IU intramuscular dose.

Ovum pick-ups were scheduled 35–36 h following hCG administration, by trans-vaginal needle aspiration under conscious sedation and ultrasound guidance. The follicular fluid was immediately transferred to the embryology laboratory for proper oocyte identification and isolation. IVF and ICSI procedures were performed as per laboratory standard protocol.

Study design

The study design was retrospective, observational and self-controlled. All fever-exposed stimulation cycles of interest (study group) were compared with the next stimulation cycle (control group) in the same women if occurring within a period of 6 months.

Outcome measures

Primary outcome measures were final number of pre-ovulatory follicles (≥16 mm) and final serum estradiol levels (pg/mL). Secondary outcome measures were total days of gonadotropin stimulation, total number of gonadotropin ampoules utilised, final number of medium-size follicles (12–15 mm), and final estradiol serum level per follicle ≥12 mm (pg/mL). Mean ratio of pre-ovulatory/total follicles ≥12 mm was also calculated.

Data collected and analysed also included patients’ age (years), primary aetiology and duration (years) of infertility.

Statistical analysis

The paired t-test was used for in-between comparison of the outcome variables of study and control cycles in women studied. P < 0.05 was considered statistically significant. Values are expressed as mean ± SD. SPSS 16.0 software was used for statistical analysis.

Results

A total of 325 stimulation cycles for IVF/ICSI were reviewed during the study period. In 25 cycles women self-reported the occurrence of febrile illnesses during the course of controlled ovarian stimulation. Six of these cycles were characterised by poor ovarian response and were considered for further analysis (study group). All febrile illnesses were reportedly due to influenza viral upper respiratory tract infections and were associated with oral temperatures exceeding 38.5°C.

The mean age of the six women studied was 30.3 (±2.2) years. Five had a history of ovulatory dysfunction: polycystic ovary syndrome (n = 4) and hypothalamic hypogonadism (n = 1). All six women had a subsequent stimulation cycle within the next 6 months (control group).

Table 1 shows the difference in ovarian follicular characteristics between fever-exposed and matched control cycles. The final number of pre-ovulatory follicles in fever-exposed cycles was significantly lower compared to matched controls (0.7 ± 0.8; 8.8 ± 2.0; p = 0.001), the final number of medium-size follicles significantly increased (21.0 ± 4.5; 10.3 ± 3.0; p = 0.001), and the final mean serum estradiol per follicle ≥12 mm significantly depressed (50.7 ± 11.7; 164.5 ± 10.9; p = 0.001). In the study group, the duration of gonadotropin stimulation was significantly longer (15.7 ± 3.3; 9.8 ± 1.3; p = 0.001), and the total number of gonadotropin ampoules utilised significantly higher compared to the matched control group (47.2 ± 10.9; 32.3 ± 4.2; p = 0.005).

Table 1.  Ovarian follicular characteristics of fever-exposed stimulation cycles compared to matched control cycles.

	Fever-exposed group (n = 6)	Control group (n = 6)	p value
Duration of stimulation (days)	15.7 ± 3.3	9.8 ± 1.3	0.001
Total number of gonadotropin ampoules	47.2 ± 10.9	32.3 ± 4.2	0.005
Final number of medium-size follicles	21.0 ± 4.5	10.3 ± 3.0	0.001
Final number of pre-ovulatory follicles	0.7 ± 0.8	8.8 ± 2.0	0.001
Final serum estradiol levels (pg/mL)	1079.2 ± 240.8	3133.33 ± 720.2	0.001
Final serum estradiol per follicle ≥12 mm (pg/mL)	50.7 ± 11.7	164.5 ± 10.9	0.001
Mean ratio of pre-ovulatory/total follicles ≥12 mm	0.03 ± 0.04	0.46 ± 0.06	0.009

All fever-exposed cycles, except for case 6, were cancelled for failure to achieve the criteria for hCG triggering of final oocyte maturation, and therefore did not reach the stage of ovum pick-up. Table 2 describes the follicular and oocyte characteristics of case 6, in which only two oocytes were collected out of 14 aspirated follicles. Both oocytes were characterised by poor quality and early degenerative changes. Both fertilised following ICSI, but did not cleave and resulted in no transfer.

Table 2.  Follicular and oocyte characteristics of fever-exposed cycle 6.

	Fever-exposed cycle 6
Ovulatory status	Polycystic ovary syndrome
Days of stimulation	19 days
Total number of gonadotropins	41 amps
Final number of medium-size follicles	12
Final number of pre-ovulatory follicles	2
Final serum estradiol	1220 pg/mL
Total number of oocytes recovered	2
Total number of zygotes	2
Total number of cleaved embryos	0
Total number of embryos transferred	0

Discussion

The findings of this observational study suggest a possible negative influence of febrile illnesses on the follicular development of women during ovarian stimulation. The reported effects were manifested in the heat-exposed cycles, and were not observed in the next stimulation cycles.

A literature search failed to identify any data reporting on the effects of hyperthermia on follicular development in humans. The veterinarian literature in contrast was found to be replete with scientific evidence linking in vivo and in vitro heat stress in animals to alterations in folliculogenesis manifested by the reduced steroidogenic capacity of theca and granulosa cells [4], [5], [8], [12], [13] and impaired follicular growth [5], [7], [14]. Although the cellular mechanisms behind these observations are not fully comprehended, they have been largely attributed to the action of heat shock proteins (HSPs). These polypeptide chain-binding proteins are believed to act as molecular chaperons to maintain protein folding and stability, protecting cells from irreversible damage during an insult [15]. HSPs were shown to submerge the cells into a state of metabolic unresponsiveness and dormancy, a mechanism that constitutes one of the cellular survival means of coping with heat stress [16–18].

The findings of this observational study suggest that hyperthermia in humans, as in animals, appears to disrupt predominantly the growth of pre-ovulatory follicles in a selective manner. In cattle and cows in vivo seasonal heat was also found to limit the degree of follicle dominance preferentially favouring the survival of more medium-size subordinate follicles [4], [19], [20]. Mouse oocytes at the dictyate stage were shown to respond to heat stress by a significantly higher synthesis of HSPs than pre-ovulatory oocytes [21]. This differential heat shock response between various stages of oogenesis increases the susceptibility of dominant follicles to hyperthermia, and could explain the differential increase in medium-size follicles and decrease in pre-ovulatory follicles in fever-exposed stimulation cycles of women.

This study also demonstrates a significant decline in estradiol production per follicle in fever-exposed cycles when matched to unexposed cycles in the same women. The effects of hyperthermia on ovarian steroidogenesis in animals are well documented. In cattle, heat stress appears to decrease estradiol levels during follicular development and interfere with the regression of the corpus luteum [5], [22]. A state of decreased responsiveness to luteinising hormone (LH) has been proposed [6], [23], and is also believed to interfere with other LH-dependent events such as loosening of the cumulus-oophorus complex within the antral space affecting the efficiency of oocyte needle aspiration. This may explain the partial ‘empty follicle syndrome’ encountered in case 6, in which only two oocytes were collected out of 14 aspirated follicles.

An evaluation of fever-exposed cycle 6 that reached ovum pick-up (Table 2) reveals a low oocyte collection yield, poor oocyte quality (early degenerative changes) and lack of embryo cleavage following fertilisation. While we believe that no universal conclusions may be made on the basis of a single observation, animal data have shown to be very useful in providing some insight into the effects of fever on oocyte developmental competence. In mice and cattle hyperthermia altered the developmental competence of follicle-enclosed oocytes in vivo and ex-vivo [9], [24] and produced morphologically and cytogenetically aberrant oocytes [25], [26] incapable of embryonic development and implantation.

It should be emphasised, however, that the extrapolation from animal data to explain the findings in human studies may not always be applicable. Although the spectrum of follicular developmental defects caused by hyperthermia has been shown to be similar in many animal species, the interpretation of findings may be complicated by species variability in both follicular dynamics and physiological body temperatures to which metabolic processes have been optimised [27]. Furthermore, while most animal studies evaluated the heat effects of controlled experimentation and increased ambient temperatures, the different nature of the heat-producing event in the present study may have rendered any comparative interpretation of data more limited. In the case of febrile illnesses, it may be difficult to separate the effects of heat alone from the confounding influence of related metabolic and inflammatory changes mediated by the viral infection [28].

This study has also a number of serious limitations including the retrospective design and the small number of observations reported. Recording of data on the basis of patients’ self-reporting introduces significant bias to the data analysed, as many events may have been simply omitted by patients’ failure to report. The detailed characteristics of the febrile events were often incomplete as to the onset and duration of fever in relation to the various stages of follicular development. Since the patient stimulation history sheets were not designed to probe specifically for a history of fever, it is possible that the detection of abnormal stimulation outcomes has triggered the search for an incriminating event in many cases, thus introducing a serious source of patient selection bias. The ovulatory status of the patients studied was not representative of the general infertile population as five of the study cohort had a particular underlying ovulatory dysfunction (polycystic ovary syndrome (PCOS) and hypothalamic hypogonadism). All of these shortcomings may have converged to obscure the actual effect of fever on follicular development and ovarian steroidogenesis in women, leading to an overestimation of the effect under investigation.

Conclusion

This comparative analysis between fever-exposed and matched controlled stimulation cycles in the same women during IVF/ICSI cycles suggests a possible negative effect of febrile illnesses on the quality of the ovarian response resulting in more cycle cancellations, nonetheless without revealing differential predisposing factors. The limitations of the study design, however, neither allow the quantification of the presumed effect, nor the characterisation of the women affected, which limits further generalisation.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.