Azoospermia: Is Sample Centrifugation Indicated? A National Survey of Practice and the Oxford Experience

1. Azoospermia: is sample centrifugation indicated? A
national survey of practice and the Oxford experience
Alexander Swanton, M.R.C.O.G., Aysha Itani, M.Sc., Enda McVeigh, M.R.C.O.G., and
Tim Child, M.D., M.R.C.O.G.
Oxford Fertility Unit, Women’s Centre, John Radcliffe Hospital, University of Oxford, Oxford, United Kingdom
Objective: Some men with azoospermia on general laboratory testing have low quantities of sperm in the ejaculate
that can only be identified through sample centrifugation and careful examination of the pellet droplets (extended
sperm preparation [ESP]). Such sperm can be used for IVF-ICSI as an alternative to either surgical sperm retrieval
(SSR) or donor sperm. The aims of the present study were to: 1) assess UK IVF clinic practice with regard to ESP
in men with azoospermia; and 2) to analyze the outcome of ESP and SSR in azoospermic men attending the
Oxford Fertility Unit.
Design: National survey of all 70 IVF units plus chart review.
Setting: Assisted conception unit.
Patient(s): One hundred twenty-two azoospermic men referred to the Oxford Fertility Unit.
Main Outcome Measure(s): Proportions of UK IVF clinics performing ESP for azoospermia. Proportions of
azoospermic men in Oxford with sperm identified at ESP and, if necessary, SSR. Relationship between serum
FSH and outcome.
Result(s): In part 1 of the study, 55 (79%) of the 70 UK IVF clinics returned completed questionnaires. Fifty
clinics (91%) routinely performed ESP for men with azoospermia on general laboratory testing, four clinics (7%)
proceeded straight to SSR without prior ESP, and one clinic varied in their approach. When clinics were asked
whether they used serum FSH levels when considering whether to proceed to SSR 28 (51%) did, 9 (16%) did not,
and 18 (33%) varied in their approach. The value placed on testicular volume similarly varied. Part 2 of the study
included 122 men referred to the Oxford Fertility Unit with azoospermia on general laboratory testing.
Eighty-seven men underwent ESP. Motile sperm was found, cryopreserved, and later used during IVF-ICSI
treatment in 19 men (22%). Eighty-one men underwent SSR (after either a negative ESP or declining ESP).
Viable sperm was found in 66 men undergoing SSR (81%). There was a statistically significant relationship
between serum FSH and the chance of retrieving sperm with SSR (Pϭ0.002) but not with ESP.
Conclusion(s): The majority (91%) of IVF clinics in the UK routinely perform ESP in men with azoospermia on
general testing. Only half routinely used serum FSH levels as predictors of SSR outcome. The value of ESP is
confirmed by our findings in Oxford. Twenty-two percent of men with azoospermia on general laboratory testing
had sufficient sperm found at ESP to proceed to IVF-ICSI without resorting to the use of either SSR or donor
sperm. Serum FSH levels were not related to the chance of finding sperm during ESP but were related to the
outcome of SSR. Our results suggest that ESP should be considered for all men with azoospermia and no apparent
obstruction. (Fertil Steril௡ 2007;88:374–8. ©2007 by American Society for Reproductive Medicine.)
Key Words: Azoospermia, extended sperm preparation, surgical sperm retrieval, FSH
Concerns have been raised over the reliability and validity of
semen analyses (1). In particular, it is important that the
diagnosis of azoospermia is as accurate as possible, because
the consequences are potentially significant. Couples with
such a diagnosis may decide to stop treatment, to consider
the use of donor sperm, or to attempt surgical sperm retrieval
(SSR) before IVF-ICSI. However, it is apparent that low
quantities of sperm may occasionally be identified using
centrifugation techniques and a thorough microscopic search
through many droplets of ejaculate sediment in otherwise
azoospermic men (extended sperm preparation [ESP]). The
likelihood of finding sperm with ESP depends on the force of
centrifugation (2, 3) in addition to the cause of azoospermia
and serum FSH concentration (4, 5). Consequently, ESP is a
potential alternative to SSR for some men with nonobstruc-
tive azoospermia on general laboratory testing (6, 7).
Approximately 40 patients are referred to the Oxford Fertility
Unit each year with the diagnosis of azoospermia. This diag-
nosis is generally made on the basis of two semen analyses
performed at least 3 months apart. Semen analyses are most
frequently performed via routine microbiology services in the
patients’ local hospital. The sample is then analyzed by the
microbiology laboratory according to their protocol. However,
not all district general laboratories have the equipment or per-
sonnel to perform a centrifuged analysis.
The aim of the present study was to analyze the figures at
our unit of men who are referred with azoospermia, subse-
Received May 25, 2006; revised and accepted November 27, 2006.
Reprint requests: Tim Child, M.A., M.R.C.O.G., Consultant Gynaecologist
& Subspecialist in Reproductive Medicine, Oxford Fertility Unit, Univer-
sity of Oxford, Level 4, Women’s Centre, John Radcliffe Hospital,
Oxford, OX3 9DU, United Kingdom (FAX: ϩ0044 (0)1865 221031; E-
mail: tim.child@obs-gyn.ox.ac.uk).
374 Fertility and Sterilityா Vol. 88, No. 2, August 2007 0015-0282/07/$32.00
Copyright ©2007 American Society for Reproductive Medicine, Published by Elsevier Inc. doi:10.1016/j.fertnstert.2006.11.121

2. quently determine their further management, and examine
current practice in other ART clinics in the UK.
MATERIALS AND METHODS
Current Practices
A questionnaire (Appendix) was sent to all 70 licensed IVF
clinics in the UK. The Medical Director (MD) was asked
under what circumstances, if any, an ESP would be used for
a man with azoospermia on repeated general laboratory
analysis. The MD was also asked if the clinic considered
serum FSH levels and testicular volumes when deciding
whether to proceed to SSR for men with azoospermia.
Extended Sperm Preparation
The charts of 122 consecutive men referred to the Oxford
Fertility Unit with azoospermia on general laboratory anal-
ysis were examined. Each man had a minimum of two
azoospermic analyses at least 3 months apart. Men were
categorized as having nonobstructive azoospermia if they
had a raised FSH level Ͼ10 IU/L. Some men had undergone
attempted reversal of vasectomy. In all the remaining cases,
no obvious cause for obstruction was noted on clinical ex-
amination. Six different hospital laboratories in the region
performed the initial basic semen analyses. Men either
elected to undergo an ESP or proceeded directly to SSR.
Those men who had azoospermia confirmed at ESP either
had SSR, no treatment, or were referred for donor insemi-
nation (DI).
The protocol for the ESP is as follows.
The sample was placed in an ungassed incubator at 37°C
to allow liquefaction and examined for the presence of
sperm. The volume and pH of the sample was recorded. If no
sperm were seen or the count was below 1 ϫ 106
, then the
whole sample was prepared using a 40% density gradient
(PureSperm; Nidacon Laboratories, Widdington, Saffron
Walden, Essex, United Kingdom).
Up to 1.5 mL semen was layered onto the gradient. This
was centrifuged at 300g for 20 minutes, after which the
seminal plasma and the top part of the gradient was removed
and discarded. A Gilson pipette was used to remove 100 ␮L
containing the pellet from the bottom of the tube. The pellet
was washed twice by placing into 4 mL and 2 mL HEPES-
buffered culture medium (MediCult Sperm Preparation; Med-
icult, Redhill, Surrey, United Kingdom) and centrifuged at
500g for 5 minutes. After the second wash the supernatant
was discarded and resuspended to an appropriate volume,
usually 50 ␮L, depending on the size of the pellet. A further
slide assessment of motility and concentration was per-
formed. If motile sperm were seen, the preparation was made
up to 400 ␮L and frozen.
If no sperm were seen by slide assessment, a more detailed
examination of the sample was made. 5 ␮L of the sperm
suspension was added to each end of 3 “long drops” under
oil. Each “long drop” was prepared by spreading 10-␮L
aliqouts of HEPES-buffered media (Medicult Sperm Prepa-
ration) in a Petri dish (Falcon 35; Becton Dickinson Lab-
ware, Oxford, United Kingdom). The drops containing the
sperm suspension were covered in paraffin oil (MediCult)
and left to settle for 10 minutes to allow any motile sperm to
swim out of the debris to the edge of the drop. The drops
FIGURE 1
Flow chart showing outcome of investigations for
122 ”azoospermic” men.
Swanton. Extended sperm preparation in azoospermia. Fertil Steril 2007.
TABLE 1
Summary of questionnaire survey data from 55 UK clinics.
Routine
ESP
FSH
used
FSH >10
IU/L
FSH >15
IU/L
FSH >30
IUL
Other
FSH
value
Testicular
volume
used
Yes 50 (91%) 28 (51%) 1 (2%) 11 (20%) 16 (29%) 18 (33%) 18 (33%)
No 4 (7%) 9 (16%) ND ND ND ND 37 (67%)
Sometimes 1 (2%) 18 (33%) ND ND ND ND ND
Note: ND ϭ no data.
Swanton. Extended sperm preparation in azoospermia. Fertil Steril 2007.
375Fertility and Sterilityா

3. were then examined for presence of sperm. If sperm were
seen, the sample was made up to 200 ␮L and frozen. If no
sperm were seen, the couple were recommended for SSR or
DI.
The SSR performed was a standard testicular sperm aspi-
ration under sedation and local anesthetic (8).
Statistical Analysis
Fisher exact test was used for the analysis of categoric data.
RESULTS
Current Practices
Fifty-five (79%) of the 70 UK IVF clinics returned com-
pleted questionnaires (Table 1). Fifty clinics (91%) routinely
performed ESP for men with azoospermia on general labo-
ratory testing. Four (7%) clinics proceeded straight to SSR
without an ESP.
In deciding whether or not to proceed to SSR, 28 clinics
(51%) routinely considered the serum FSH concentration
(Table 1). Nine clinics (16%) did not consider the FSH level
at all, and 18 (33%) varied in their approach. An SSR was
not recommended by clinics if the FSH level was elevated,
with 1 clinic (2%) using Ͼ10 IU/L, 11 clinics (20%) using
Ͼ15 IU/L, and 16 clinics (29%) using Ͼ30 IU/L. Twelve
clinics (22%) stated that they used a level other than these
three options and commented that a specific level would not
necessarily be used but would enable greater prediction of
success.
The value placed on testicular volume also varied among
clinics (Table 1). When questioned whether testicular vol-
ume was used as an indication to perform SSR or not, only
18 clinics (33%) stated that they did. The particular testicular
volume used ranged from 2 to 10 mL, but many stated that
it was subjective.
Extended Sperm Preparation
One hundred twenty-two men referred to the Oxford Fertility
Unit with azoospermia on general laboratory testing were
included in the second part of the study. Of these patients, 32
(26%) had had a reversal of vasectomy but were diagnosed
as azoospermic on subsequent semen analysis.
Figure 1 shows the outcome of the 122 azoospermic men
undergoing ESP and/or SSR. Eighty-seven men elected to
have an ESP. Table 2 shows the outcome of ESP in these 87
patients: 21 had had vasectomy reversal (FSH not routinely
done); in 14 serum FSH was not measured; and in 52 serum,
FSH was not related to the chance of finding sperm (Fisher
exact test: PϭNS). Overall motile sperm was found, cryo-
TABLE 3
Outcome of 81 surgical sperm removal (SSR) patients.
Sperm found
at SSR
FSH not
performed
FSH <10
IU/La
FSH 10–14.9
IU/La
FSH 15–19.9
IU/La
FSH 20–29.9
IU/La
FSH ≥30
IU/La
Yes 29 (91%) 24 (89%) 9 (90%) 2 (40%) 2 (29%) 0
No 3 (9%) 3 (11%) 1 (10%) 3 (60%) 5 (71%) 0
Total 32 27 10 5 7 0
a
Fisher exact test: Pϭ.002.
Swanton. Extended sperm preparation in azoospermia. Fertil Steril 2007.
TABLE 2
Outcome of 87 patients who had extended sperm preparation (ESP).
Sperm found
at ESP VR
No FSH
level
performeda
FSH <10
IU/Lb
FSH 10–14.9
IU/Lb
FSH 15–19.9
IU/Lb
FSH 20–29.9
IU/Lb
FSH ≥30
IU/Lb
Yes 1 (5%) 4 (29%) 6 (26%) 5 (36%) 1 (20%) 2 (33%) 0 (0%)
No 20 (95%) 10 (71%) 17 (74%) 9 (64%) 4 (80%) 4 (66%) 4 (100%)
Total 21 14 23 14 5 6 4
Note: VR ϭ vasectomy reversal.
a
Excluding patients with vasectomy reversal.
b
Fisher exact test: PϭNS.
Swanton. Extended sperm preparation in azoospermia. Fertil Steril 2007.
376 Swanton et al. Extended sperm preparation in azoospermia Vol. 88, No. 2, August 2007

4. preserved, and later used during IVF-ICSI treatment in 19
men (22%).
Eighty-one men underwent SSR, either following a failed
ESP or by choosing not to have an ESP (Fig. 1 and Table 3).
The serum FSH level was highly significantly related to the
chance of finding sperm at SSR (Fisher exact test: Pϭ0.002).
The data showed that with increasing FSH values, there is a
reduced chance of retrieving sperm at SSR (Table 3). The
majority of the 32 men who did not have serum FSH checked
were those with a history of vasectomy reversal. Overall,
viable sperm was found, cryopreserved, and later used dur-
ing IVF-ICSI treatment in 66 men (81%).
DISCUSSION
The present U.K. national survey demonstrates that the ma-
jority of assisted conception clinics perform ESP in
azoospermic men. This approach is supported by data from
our own unit in that ESP identified sperm in 22% of
azoospermic patients. However, the chance of finding sperm
using ESP in men with azoospermia after attempted vasec-
tomy reversal was lower, at 5%. Indeed, in the 66 men with
no history of vasectomy and reversal, the chance of finding
sperm at ESP was 30% (20 out of 66). Multiple samples
were not analyzed, because two basic semen analyses plus an
ESP had been performed. If the ESP was unsuccessful, men
were counseled to move on to SSR, consider DI, or stop
treatment. A further ESP was not recommended in this
situation. If sperm are identified, further samples are some-
times taken to increase the number of sperm for subsequent
treatment.
Although semen parameters fluctuate, we feel that per-
forming multiple pellet analyses would not change the man-
agement of the patient.
There have been previous reports with smaller patient
cohorts (Table 4). Ron-El et al. (6) used ESP for 49 men with
azoospermia and identified the presence of sperm in 17
(35%). In a further study, Timm et al. (7) analyzed the
presence of spermatozoa and spermatids in the ejaculate of
27 men with nonobstructive azoospermia. Initially only one
sample was analyzed by centrifugation, and if that was
confirmed as azoospermic then a further sample was pro-
duced. This was repeated if the second sample was also
azoospermic. The results are similar to our findings, with
37% of men having viable sperm identified on ESP. No
sperm were identified in any of the men undergoing their
third sample, but in this situation Timm et al. identified
spermatids in the ejaculate of 41% of cases. Identification of
spermatids was not performed in the present study, but it
could be a potential option for patients in the future.
The presence of spermatids is possibly more likely in men
with azoospermia, and, although there have been successful
outcomes (9, 10), the use of spermatids in ICSI-IVF treat-
ment should still be considered experimental (11, 12).
The place of serum FSH levels in predicting outcome is
not so clear. In particular, the present survey demonstrated
that only half of clinics routinely used the serum FSH level
as a predictor of success with SSR. In men with elevated
FSH levels, even in the presence of reduced testicular vol-
ume, a number of studies have shown successful sperm
retrieval (13–15). Raman and Schlegel (16) concluded that
spermatozoa recovery was independent of testicular volume
and serum FSH level, which is consistent with other reports
(17). We found the serum FSH level to be predictive of
outcome for SSR but not for ESP. We did not find a cut-off
value above which an SSR was not to be recommended.
Patients are informed of a reduced chance of successful
sperm retrieval with an increasing FSH value but not de-
clined treatment.
In conclusion, the present results suggest that ESP should
be considered for all men diagnosed with azoospermia on
general testing where there is no apparent obstruction. Serum
FSH levels are not predictive of ESP outcome, because
sperm may be found even with apparently severe testicular
failure and FSH levels Ͼ20 IU/L. However, serum FSH
levels are predictive of outcome for SSR, although only 51%
of U.K. clinics routinely use FSH to select patients.
REFERENCES
1. Keel BA. How reliable are results from the semen analysis? Fertil Steril
2004;82:41–4.
2. Corea M, Campagnone J, Sigman M. The diagnosis of azoospermia
depends on the force of centrifugation. Fertil Steril 2005;83:920–2.
3. Niederberger C. The diagnosis of azoospermia depends on the force of
centrifugation. J Urol 2005;174:2300.
TABLE 4
Summary of findings of extended sperm preparation (ESP) in azoospermic men.
Present study Ron et al. 1997 Timm et al. 2005
Sample size, n 87 49 27
Spermatozoa seen, n (%) 19 (22%) 17 (35%) 10 (37%)
Spermatids seen, n (%) N/A N/A 11 (41%)
Note: N/A ϭ not available.
Swanton. Extended sperm preparation in azoospermia. Fertil Steril 2007.
377Fertility and Sterilityா

5. 4. Ezeh UI. Beyond the clinical classification of azoospermia: opinion.
Hum Reprod 2000;15:2356–9.
5. Schlegel PN. Causes of azoospermia and their management. Reprod
Fertil Dev 2004;16:561–72.
6. Ron-El R, Strassburger D, Friedler S, Komarovski D, Bern O, Soffer Y, et
al. Extended sperm preparation: an alternative to testicular sperm extraction
in nonobstructive azoospermia. Hum Reprod 1997;12:1222–6.
7. Timm O Jr., Cedenho AP, Spaine DM, Buttignol MH, Fraietta R, Ortiz V, et
al. Search and identification of spermatozoa and spermatids in the ejaculate of
nonobstructive azoospermic patients. Int Braz J Urol 2005;31:42–8.
8. Gorgy A, Podsiadly BT, Bates S, Craft IL. Testicular sperm aspiration
(TESA): the appropriate technique. Hum Reprod 1998;13:1111–3.
9. Kahraman S, Polat G, Samli M, Sozen E, Ozgun OD, Dirican K, et al. Multiple
pregnancies obtained by testicular spermatid injection in combination with
intracytoplasmic sperm injection. Hum Reprod 1998;13:104–10.
10. Mansour RT, Fahmy IM, Taha AK, Tawab NA, Serour GI, Aboulghar
MA. Intracytoplasmic spermatid injection can result in the delivery of
normal offspring. J Androl 2003;24:757–64.
11. Levran D, Nahum H, Farhi J, Weissman A. Poor outcome with round
spermatid injection in azoospermic patients with maturation arrest.
Fertil Steril 2000;74:443–9.
12. Ghazzawi IM, Alhasani S, Taher M, Souso S. Reproductive capacity of
round spermatids compared with mature spermatozoa in a population of
azoospermic men. Hum Reprod 1999;14:736–40.
13. Hibi H, Ohori T, Yamada Y, Honda N, Asada Y. Probability of sperm
recovery in nonobstructive azoospermic patients presenting with testes
volume less than 10 mL/FSH level exceeding 20 mIU/mL. Arch Androl
2005;51:225–31.
14. Jezek D, Knuth UA, Schulze W. Successful testicular sperm extraction
(TESE) in spite of high serum follicle stimulating hormone and
azoospermia: correlation between testicular morphology, TESE results,
semen analysis and serum hormone values in 103 infertile men. Hum
Reprod 1998;13:1230–4.
15. Kim ED, Gilbaugh JH, third ??, Patel VR, Turek PJ, Lipshultz LI.
Testis biopsies frequently demonstrate sperm in men with azoospermia
and significantly elevated follicle-stimulating hormone levels. J Urol
1997;157:144–6.
16. Raman JD, Schlegel PN. Testicular sperm extraction with intracyto-
plasmic sperm injection is successful for the treatment of nonobstruc-
tive azoospermia associated with cryptorchidism. J Urol 2003;170:
1287–90.
17. Tournaye H, Verheyen G, Nagy P, Ubaldi F, Goossens A, Silber S,
et al. Are there any predictive factors for successful testicular sperm
recovery in azoospermic patients? Hum Reprod 1997;12:80–6.
APPENDIX
Clinic Questionnaire
Please tick appropriate answer.
1. Does your unit routinely perform a detailed semen
analysis (e.g., andrology assessment or extended sperm
preparation) on patients diagnosed as azoospermic in gen-
eral laboratories? Yes Œ No Œ Sometimes Œ
2. If no, do you proceed directly to surgical sperm retrieval
(SSR)? Yes Œ No Œ Sometimes Œ
3. Do you use serum FSH levels to guide whether SSR is
appropriate? Yes Œ No Œ Sometimes Œ
4. If yes, above what FSH level would you not perform SSR?
(Please ring answer.) Ͼ10 IU/L Ͼ15 IU/L Ͼ30 IU/
L Other
5. Do you use testicular volume to guide whether an SSR is
appropriate? Yes Œ No Œ Sometimes Œ
6. If yes, what volume do you use to indicate whether or not
to perform an SSR?
378 Swanton et al. Extended sperm preparation in azoospermia Vol. 88, No. 2, August 2007