Prioritising for fertility treatments—the effect of excluding women with a high body mass index
Dr WR Gillett, Department of Obstetrics and Gynaecology, University of Otago, Dunedin, New Zealand. Email firstname.lastname@example.org
The effect of clinical priority access criteria for access to infertility treatment was examined for women outside the body mass index (BMI) range of 18–32 kg/m2. Treatments and outcomes were analysed from 1280 cases referred from 1998 to May 2005. Sixteen percent of women had a BMI of >32 kg/m2. Overall, 38% of these women had a birth from conceiving a treatment-related pregnancy or spontaneous pregnancy, compared with 52% of women with BMI < 32 kg/m2. Weight loss allowed women in the BMI group >32<35 kg/m2 to access treatment, but women in higher BMI groups were less successful.
In New Zealand (NZ), ranking patients for elective, publicly funded procedures has been achieved by the development of clinical priority access criteria (CPAC), being first introduced in the mid-1990s for several procedures.1 A CPAC for patients with infertility seeking assisted reproduction technique was implemented in 2000.2 The assisted reproduction technique CPAC was to provide a rationing basis for public access to treatment for couples who were most in need but balanced by those who would benefit most from treatment.3
Seven separate criteria were developed for the CPAC, each having subcategories for which points were weighted, reflecting the need and also the benefit obtained from assisted reproduction technique treatment.2 The seven criteria included: a prognostic score derived from the infertility diagnosis and reflecting the probability of spontaneous pregnancy, the woman’s age, the woman’s basal follicle-stimulating hormone level, whether the woman smoked or not, the duration of the couple’s infertility, the number of children and whether either partner was sterilised or not. The prognostic score was based on subcategories of the main causes for infertility: ovulation disorders, endometriosis, other tuboperitoneal disorders and male factor.2 The CPAC score would increase as the duration of infertility became longer (reaching a maximum by ≥5 years) or would decrease if the woman reached 40 years of age. The highest score possible was 100 points and the threshold for access to publicly funded assisted reproduction technique was set at ≥65 points. This meant that the couples with the worst prognosis for spontaneous pregnancy could access assisted reproduction technique after 1 year of duration of their infertility (providing other criteria were favourable).
Because obesity was considered to reduce a woman’s chance of successful assisted reproduction technique, the CPAC could only be applied to women who were inside the body mass index (BMI) range of 18–32 kg/m2. Women outside this range were only accepted on the basis that they had undergone weight improvement to within the agreed range. In 1998, the Otago Fertility Service (OFS) of the Otago District Health Board, based in the university city of Dunedin was the first to use the CPAC tool, on a trial basis, before it became widely accepted in NZ from the beginning of 2000.
The aim of this study was to assess the effect of the new CPAC prioritisation on women outside the BMI range of 18–32 kg/m2. To do this, all referred cases were scored by the CPAC, irrespective of their BMI and whether they had elected to have treatment. Treatments and outcomes were analysed, across a range of BMI subcategories, from all cases referred to the OFS from 1998 to May 2005.
Only couples being referred for the first time were included in the analysis. Of these 1365 women, the BMI was known in 1280 cases. All were scored by the CPAC tool at entry and recalculated at yearly intervals. The score, though, was not relevant unless that couple elected to have one of the assisted reproduction technique treatments or were within the BMI range of 18–32 kg/m2. The treatments requiring a CPAC score of ≥65 were all procedures using in vitro fertilisation (IVF), artificial insemination by intrauterine insemination (IUI), donor insemination (DI) or ovulation induction using gonadotrophins. Treatments that were not included in the CPAC were ovulation induction with clomiphene citrate, weight improvement programmes and surgery. Surgery included a range of reconstructive microsurgical operations (including tubal reanastomosis and vasectomy reversal) and endoscopic procedures.
Since many couples had a number of different types of treatment, the last planned (although not necessarily started) treatment was regarded as the reference treatment relating to outcome. For assisted reproduction technique treatments, the number of commenced cycles was recorded, and for IVF, the number of fresh and thaw embryo transfers was recorded. The number of nonassisted treatments (e.g. cycles of ovulation induction) was not recorded.
Couples scoring ≤65 points, and electing to have one of the assisted reproduction technique treatments, were placed on an active review list that allowed time for spontaneous conception while waiting for the score to change to ≥65. Women outside the BMI range of 18–32 kg/m2 and who had a score ≥65 were also placed on active review to allow time for weight improvement. All of these women were offered a referral to a dedicated dietician. The size of any weight change, however, was not recorded. In some cases, where significant weight loss had not resulted in the woman reaching the BMI threshold of ≤32 kg/m2, acceptance for assisted reproduction technique was considered, providing the woman was in the >32<35 kg/m2 range. This discretionary authority is known as clinical override, and fertility clinics were allowed to use this in only 5% of all publicly funded assisted reproduction technique cycles.
The infertility clinic in the hospital is the only source of secondary and tertiary care for infertility in the region and both public and private treatments were offered. Some couples on active review elected to have private treatment, although women with a BMI outside the range of 18–32 kg/m2 were still encouraged to seek weight improvement.
The end point for each case was either the date a woman was known to have conceived (providing the pregnancy continued to birth), the date of withdrawal from the programme or, for cases lost to follow up, the last date seen at the clinic. The outcomes that were measured included both spontaneous and treatment-related pregnancies that resulted in birth. Miscarriages were not considered in the pregnancy data. Other end points were elective withdrawals from the programme after completed failed treatment and a composite group of withdrawals for other reasons (withdrawing without undergoing treatment, couples who separated or were lost to follow up).
All outcomes were evaluated across five BMI categories (<18, 18–32, >32<35, 35<40 and ≥40 kg/m2). Categorical variables used the chi-square test and comparisons of mean values used analysis of variance with post hoc comparisons by Duncan’s multiple range test (α= 0.05).
One hundred and twenty women withdrew within 1 year and without treatment, most of them electing not to proceed after the initial assessment. There was no difference in withdrawal rate between the five BMI categories. Of the remaining 1160 women, 180 (16%) had a BMI > 32 kg/m2. Ninety percent were of European background and 5% were Maori or Pacific Islanders. The mean BMI (kg/m2± SD) of these two groups was 26.3 ± 5.8 and 28.8 ± 7.3, respectively (P= 0.003). The mean age (±SD) at referral was 32.9 ± 5.0 years, with no statistically significant differences between the BMI groups. The mean duration of infertility was 2.7 ± 2.3 years, with no statistically significant differences between the groups. However, when all women with BMI > 32 kg/m2 were grouped together, there was a statistically significant difference in duration (3.1 years compared with 2.6 years for women with BMI ≤ 32 kg/m2, P= 0.006).
First assessment CPAC scores were similar in each of the groups (Table 1). Sixty-one percent of cases had at least one moderately severe diagnostic factor (either an ovulation disorder, endometriosis, other tuboperitoneal or male factor). When these data included moderately severe ovulation disorders only (amenorrhoea or oligomenorrhoea), there was the expected increase in numbers in this diagnostic group (Table 1).
Table 1. Characteristics and outcome by BMI groups for women referred with infertility (percentage in brackets)
|Number of women||15||965||77||56||47||1160|
|Mean CPAC score at referral (±SD)||56 ± 36||51 ± 27||58 ± 17||58 ± 28||51 ± 32||52 ± 28|
|Number of women with at least one moderately severe diagnostic category (χ2= 9, df= 4, P= 0.05)||9 (60)||579 (60)||55 (71)||40 (71)||23 (49)||706 (61)|
|Number of women with at least moderately severe ovulation disorder (χ2= 32, df= 4, P < 0001)||3 (20)||108 (11)||18 (23)||17 (30)||13 (27)||159 (14)|
|Couples with CPAC score ≥ 65 by after 1 year (χ2= 6.9, df= 4, P= 0.14)||8 (53)||464 (48)||47 (61)||33 (59)||23 (49)||575 (50)|
|Number of women having publicly funded assisted reproduction technique (χ2= 23, df= 4, P < 0.0001)||6 (40)||314 (33)||28 (36)||10 (18)||2 (4)||360 (31)|
|Number of women having private assisted reproduction technique only (χ2= 16, df= 4, P= 0.04)||2 (13)||149 (15)||4 (5)||2 (4)||2 (4)||159 (14)|
|Last planned treatment (number of cases and births from spontaneous and treatment-related pregnancies)|
| All births||2||190 (43)||11 (37)||5 (29)||0||208 (41)|
| All births||0||61 (50)||2 (25)||2 (50)||0||65 (46)|
| All births||0||50 (61)||9 (60)||7 (64)||3 (50)||69 (60)|
|Gonadotrophins|| ||16||2||1|| ||19|
| All births|| ||12 (75)||1 (50)||0|| ||13 (68)|
|Weight loss|| ||3||5||16||31||55|
| All births|| ||0||2 (40)||3 (18)||8 (25)||13 (24)|
| All births||0||48 (44)||4 (44)||1 (25)|| ||53 (43)|
|Other*|| ||16||1|| ||17|
| All births|| ||5 (31)||0|| ||5 (29)|
|Number of women conceiving with treatment (χ2= 16.4, df= 4, P= 0.002)||2 (13)||306 (32)||22 (29)||11 (20)||4 (9)||345 (30)|
|Number of women conceiving spontaneously (χ2= 3.0, df= 4, P= 0.55)||1 (7)||202 (21)||13 (17)||10 (18)||8 (17)||234 (20)|
|Number of women completing treatment without pregnancy (χ2= 3.9, df= 4, P= 0.42)||5 (33)||199 (21)||19 (25)||16 (29)||11 (23)||250 (22)|
|Number of women still in treatment (χ2= 14.5, df= 4, P= 0.006)||6 (40)||194 (20)||17 (22)||14 (25)||19 (40)||250 (22)|
|Number of women lost to follow up (χ2= 1.6, df= 4, P= 0.82)||1 (7)||64 (7)||6 (8)||5 (9)||5 (11)||81 (7)|
Although similar proportions in each BMI category were eligible for publicly funded assisted reproduction technique, the number actually receiving public treatment was lower in the higher BMI groups. Furthermore, the women with BMI > 35 kg/m2 were more likely to be still in the programme, either on active review or in treatment programmes. Women with BMI > 32 kg/m2 were also less likely to receive private assisted reproduction technique.
Table 1 describes the last intended treatment for each woman and gives the number of births resulting from either the treatment or a spontaneous pregnancy that resulted independently from that treatment. Not all of these intended treatments were started. For example, of the 504 planned IVF cases, 380 women commenced IVF, resulting in 174 births. The mean number of IVF cycles commenced for these 380 women was 1.7 ± 1.1, with a mean of 2.5 ± 1.7 embryo transfers (fresh and frozen embryos). Another 34 women conceived before the planned IVF was started. The mean (±SD) number of IUI/DI treatments was 2.1 ± 1.4. There were 183 women who did not have a planned treatment, 153 of them conceived spontaneously. When weight loss treatments were the last planned treatment, few women achieved a pregnancy by this measure alone. These data, however, do not record women who conceived by other treatments, who may also have had earlier weight improvement programmes.
All treatment-related pregnancies and spontaneous pregnancies resulting in births are summarised in the final outcomes of the Table 1. Two hundred and thirty-four women conceived spontaneously and had a birth in the follow-up period, with a mean conception interval of 0.69 ± 0.72 years from referral. Despite similar spontaneous pregnancy rates across the BMI categories, the women in the higher BMI categories had a significantly reduced treatment birth rate. Overall, 68 (38%) and 511 (52%) of the women >32 kg/m2 and ≤32 kg/m2, respectively, had a birth after conceiving spontaneously or with treatment.
This report aimed to highlight the challenges of providing fertility treatment to women who are severely overweight. The NZ CPAC development was about providing a mechanism for rationing limited public funds. It was felt that women with a low or high BMI were likely to be less responsive to treatment, and for the CPAC score to count towards access, the BMI had to be within the accepted range. Since there were few quantitative data available, the BMI was not included in the scoring system itself. These data have since become available that justifies the importance of weight loss in the initial management.4,5 Our data from the <18 kg/m2 group were too few for meaningful interpretation.
In this study, women in the >32<35 kg/m2 group were able to modify their lifestyle and achieve a weight loss, enabling access to treatment. Some of these women would have gained access even though their BMI had not reached the threshold of 32 kg/m2. A significant limitation of this analysis was that the actual weight loss was not recorded. Despite this, their reproductive outcomes were mostly comparable to those in the accepted BMI range. The assumption is that this group, being close to the accepted range for assisted reproduction technique, was more motivated to achieve a weight reduction. This did not apply to the women with a BMI of ≥35 kg/m2, their reproductive outcomes being significantly reduced, although again any weight reduction was not recorded. Our data were also limited in that it did not record those who had participated in weight improvement programmes; it recorded only those intending to participate. Furthermore, these programmes comprised mainly dietary advice and supervision by a hospital dietician or by self-directed measures. All other obesity treatments were provided on a limited ad hoc basis. These included lifestyle changes and treatments involving exercise, behaviour modification and pharmacological treatments but did not involve group activities. A group environment particularly enhances reproductive outcome.4,5 Group activities are available in other NZ centres and so our experience coming from only one district health board (representing only 10% of all fertility treatments offered in NZ) may underestimate the effect in women with a higher BMI.
Although there was a good outcome in the BMI group of >32<35 kg/m2, our study does not show if this improvement was necessarily related to weight reduction. We consider that most evidence supports the need to use weight improvement measures, but the threshold at which this should take precedence is unclear.5 In NZ, the BMI threshold of ≤32 kg/m2 seems harsh for Maori or Polynesian women. The BMI level of 32 kg/m2 for a 30-year-old European woman is at about the 90th percentile compared with 70th percentile for Maori and Pacific Island women.6 Despite this, the same CPAC model and the BMI threshold was used for all ethnic groups.
We were encouraged by the good reproductive performance of those close to the BMI threshold and the compliance at which many women took up the task to gain access to treatment programmes. However, women in the higher BMI ranges still provide challenges and from our experience will still be disadvantaged from accessing infertility treatments without some other way to achieve weight loss. The NZ system has not promoted public funding to help these women, but given their poor outcome, it is timely that weight reduction programmes should be a major focus in the future of prioritised health care.