Laboratory diagnosis, molecular characteristics, epidemiological and clinical features of an outbreak of measles in a low incidence population in Australia

Journal of Clinical Virology 54 (2012) 168–173

Contents lists available at SciVerse ScienceDirect

Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv

Laboratory diagnosis, molecular characteristics, epidemiological and clinical features of an outbreak of measles in a low incidence population in Australia Jude Jayamaha a , Philippa L. Binns b , Michael Fennell a , Mark J. Ferson b,e , Peter Newton c , Thomas Tran d , Michael Catton d , Peter Robertson a , William Rawlinson a,∗ a

Virology Division, Department of Microbiology, South Eastern Area Laboratory Services, Prince of Wales Hospital, Sydney, Australia South Eastern Sydney Illawara Public Health Unit, Randwick, New South Wales, Australia c Department of Microbiology, South Eastern Area Laboratory Services, Wollongong Hospital, Wollongong, Australia d Victorian Infectious Diseases Reference Laboratory, Melbourne, Australia e School of Public Health & Community Medicine, University of New South Wales, Kensington, Australia b

a r t i c l e

i n f o

Article history: Received 10 November 2011 Received in revised form 11 February 2012 Accepted 24 February 2012 Keywords: Measles Laboratory diagnosis Complications Genotypes

a b s t r a c t Background: Prompt and accurate laboratory diagnosis of measles is essential for case detection, outbreak management and ongoing surveillance in low incidence countries. Several disease markers are employed for diagnosis and are important to determine epidemiological and molecular characteristics for future control measures. Objectives: To report different disease markers, genotypes and epidemiology of a measles outbreak in Australia, a low incidence country. Study design: A retrospective descriptive study of the clinical and epidemiological features and laboratory diagnosis in 16 confirmed measles cases using measles serum IgM/IgG, antigen detection (IFA), viral RNA detection by real-time PCR and genotyping results for respiratory and urine specimens processed in one reference laboratory. Results: Of the 16 confirmed measles cases, 11 were young adults aged between 20–35 years and 15 were not age-appropriately vaccinated. The most common genotype detected was D9 (11/16), followed by D4 (1/16) and D8 (1/16). Two imported cases were from the Philippines (D4) and Italy (D9). Of six disease markers, respiratory swab PCR and serum IgM gave the highest percentage (100%) of positive samples for confirmed cases followed by urine PCR (90.9%), serum PCR (66.6%), urine IFA (54.5%) and respiratory IFA (46.2%). Conclusions: Measles should be considered in the differential diagnosis of a presentation with fever and rash, even in countries in the elimination phase of measles control. Genotyping is a powerful molecularepidemiological tool to assist low incidence countries towards eradication goals. Improving vaccination coverage remains essential, particularly in young adults and travellers. © 2012 Elsevier B.V. All rights reserved.

1. Background Measles is a highly communicable viral disease that may lead to serious complications.1 The incidence is low in many developed countries including Australia. However, since late 2009, many outbreaks in several developed countries especially in Europe have occured.2 Since 2000, measles notifications for Australia have ranged from <1 to 9 cases per million per year (range 10–190 per year, average 82 per year); in 2011, 190 cases were notified.3 To achieve and maintain elimination status, measles must be

∗ Corresponding author at: Virology, SEALS Microbiology, Level 4 Campus Centre, Prince of Wales Hospital, Randwick, NSW 2031, Australia. Tel.: +61 2 93829113; fax: +61 2 93984275. E-mail address: [email protected] (W. Rawlinson). 1386-6532/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.jcv.2012.02.025

confirmed or excluded as a matter of urgency by specific testing for all patients presenting with a fever and maculopapular nonvesiculating rash.4 Accurate and prompt laboratory diagnosis is essential for the public health response. Diagnosis is based on measles IgM detection in serum, measles antigen detection, viral RNA detection and, less frequently, virus isolation.5 Genotyping of measles virus (MV) allows the origin of measles importations to be traced, particularly in countries that have reached the elimination phase, such as Australia.4

2. Objectives To define the epidemiology and molecular epidemiology of a measles outbreak that occurred on the east coast of New South

Table 1 Immunization, contact history, genotype and disease outcome of confirmed cases of measles. Case no.

Date of onset of rash

Immunization history

Contact history/link

Cluster number

Genotype

Hospitalized/duration

F/17

16/2/11

D4

No

F/21

01/3/11

Travelled in Italy during exposure period Case 1 (sister)

1

2

1

N/A

No

3

F/30

03/3/11

Not vaccinated – parent conscientious objector Not vaccinated – parents conscientious objector. Vaccinated as a contact UNK

2

a

No

4 5

M/30 M/31

15/3/11 05/3/11

UNK UNK

2 3

D9 D9

Yes – 3 days and 8 days Yes – 3 days

6 7 8

M/29 F/33 M/3

20/3/11 23/3/11 11/3/11

UNK UNK Not vaccinated – contraindications

3 3 4

D9 D9 D9

No No Long term inpatient

9

F/1.2

22/3/11

4

D9

No

10

M/1.1

11/3/11

4

a

No

11

F/10

21/3/11

Not vaccinated, was 2 months overdue for the 1st dose Not vaccinated, was due for the 1st dose but developed measles Not vaccinated – contraindications

4

D9

b

12

F/34

20/3/11

UNK

4

D9

Yes – 4 days

13

F/26

10/2/11

5

D9

Yes

14

M/35.

06/3/11

Not vaccinated as not provided in country of origin UNK

Unknown source case. Visited ED during exposure period Case 3 (wife) Unknown source case/epi link Case 5 Case 5 Unknown source case/epi link, but on paediatric ward at time of exposure Case 8 (brother) and visited paediatric ward during exposure Unknown source case but visited paediatric ward during exposure Contact of case 10 (brother) and inpatient of paediatric ward Contact of case 10 (son) and visited paediatric ward Travelled in Philippines during exposure period

5

D9

No

15

M/34

17/4/11

UNK

–

D9

Yes – 6 days

16

M/18

23/4/11

Yes, 2 doses

–

D8

No

Travelled to work with case 13 on a few occasions Unknown source case/epi link Attended University where other cases were known to have attended

Complications

Encephalitis

Pregnant

J. Jayamaha et al. / Journal of Clinical Virology 54 (2012) 168–173

Sex/age (years/months)

1

Pneumonitis, very high LFT

UNK, immunization status not known by case or doctor; N/A, not available. a PCR negative at reference laboratory. b Child had many admissions due to multiple medical and congenital conditions.

169

170

J. Jayamaha et al. / Journal of Clinical Virology 54 (2012) 168–173

Table 2 Disease markers in confirmed cases (1–16) and excluded cases (17–34) of MV infection. No.

Resp. swab/NPA PCR

Resp. swab/NPA IFA

Urine PCR

Urine IFA

IgM

IgG

1 2 3 4a 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

N/A N/A + + + + + + + + + + + + N/A + − − − N/A − −

N/A − − − + + − N/A + + + + − + N/A − + − + N/A N/A +

+ N/A + + + + + N/A − − + + N/A + + + N/A − − N/A N/A −

N/A N/A − + + + − N/A IND N/A N/A + N/A + − + N/A + − N/A N/A +

+ + + + + + + + +a N/A + + + N/A + + N/A − − + + N/A

+ + − − − − − − − N/A − − − N/A N/A + N/A + − − − N/A

PCR, polymerase chain reaction; IFA, indirect fluorescent assay; N/A, not available; IND, indeterminate. In cases 23–34 all available investigations were negative except measles IgG was positive in 7 cases. a CSF sample negative for MV real time PCR.

Wales during February to May 2011 and to compare utility of different laboratory markers of infection. 3. Study design Respiratory (nasopharyngeal aspirate or throat/nasal swabs), urine and serum specimens from suspected and confirmed cases of measles received at South Eastern Area Laboratory Services (SEALS) laboratories from 1st February 2011 to 31st May 2011 were analyzed using the methods described below. Demographic details, epidemiological links to suspected or confirmed measles cases, clinical data, immunization and travel history of patients were obtained retrospectively from clinical and public health surveillance records (Table 1). Confirmed cases were defined in accordance with National Notifiable Diseases Case Definitions.6 After public health follow up, cases suspected to be measles by clinicians were excluded when clinical evidence was not compatible with measles case definition (e.g. no fever at rash onset, or rash with no coryza, cough or conjunctivitis) and/or laboratory investigations for measles were negative. When one or more laboratory tests for measles were positive, more specific disease markers (PCR and specific IgM), detailed clinical and epidemiological histories were taken into consideration to further exclude suspected cases (Table 2). 3.1. Laboratory methods 3.1.1. Antigen detection Respiratory and urine samples were processed and stained using an indirect fluorescent (IFA) assay (Argene, USA) as previously described.7 3.1.2. Serology assays Sera were tested for the presence of MV-specific IgM (Chorus, Diesse Diagnostics, Italy – evaluated against Behring EIA and immunofluorescence, sensitivity of 100% and specificity of 98.1%) and IgG (NovaTec, Immunodiagnostica, Germany, sensitivity & specificity >95.5%) according to the manufacturers’ instructions. These assays regularly give correct results in the international quality assurance programme for measles antibody.

3.1.3. MV RNA detection using real-time PCR RNA was extracted from 200-␮l of respiratory, urine and serum samples using either Magna Pure total nucleic acid isolation kit (Roche Diagnostics) or Siemens sample prep 1.0 on Kingfisher flex. With each extraction batch a known positive sample was included. Real time RT-PCR (rt-PCR) was performed using a CE certified commercial assay (Liferiver, Shanghai ZJ Bio-Tech, China states a sensitivity of: 5 × 103 copies/ml). In brief, 5 ␮l of purified RNA extract, positive and negative controls were added to 15 ␮l of master mix, and subjected to 40 cycles of 45 ◦ C for 10 min, 1 cycle; 95 ◦ C for 15 min, 1 cycle; 95 ◦ C for 5 s and 60 ◦ C for 30 s. Results were interpreted only when the internal control amplification signal was present for each sample. Samples that did not exhibit an internal control signal were re-extracted and retested unless positive. Above tests were performed in one laboratory (SEALS) except for case 1 (urine PCR and serology at VIDRL, urine IFA at South Coast Independent Laboratory), case 2 (swab PCR at Institute for Clinical Pathology and Medical Research – Sydney, serology at Symbion Laverty), cases 3, 6, 9 and 14 (serology at Southern IML Pathology/DHM), case 13 (serology at South West Area Health Service). All laboratories used Therapeutic Goods Administration licensed and National Association of Testing Authorities approved Enzyme immuno assays and IFA tests quality assured by Royal College of Pathologists of Australia – Quality Assurance Programme. 3.1.4. Genotyping of MV Molecular characterization was performed at Victorian Infectious Diseases Reference Laboratory (VIDRL), Melbourne to determine the genotype of MVs involved sequencing 450 nucleotides coding for the COOH terminus of the viral nucleoprotein (N) gene as previously described.8 The MV sequences were submitted to GenBank with accession numbers JQ609263, JQ609264, JQ609265, JQ609266, JQ609267, JQ609268, JQ609269, JQ609270, JQ609271, JQ609272, JQ609273, JQ609274, JQ609275. 4. Results Of 34 patients investigated for suspected measles, 16 were confirmed to have measles based on laboratory, clinical and

J. Jayamaha et al. / Journal of Clinical Virology 54 (2012) 168–173

171

Fig. 1. Phylogenetic tree comparing 13 measles viruses and reference strains detected on the east coast of New South Wales, Australia, between February and May 2011. The distance-based phylogenetic tree was constructed by neighbour-joining methods using the F84 model of evolution with optimized parameters based on 456nt nucleoprotein sequences. The genotype A reference strain was used as an outgroup to root the tree. Bootstrap values greater than 90% are indicated. Measles virus reference strains are indicated in italics.

epidemiological results (Table 2). The mean age for confirmed cases was 22.4 ± 12.39 (SD) years (range 1–35 years). A quarter (4/16) of confirmed cases were less than 11 years; the remainder were young adults (Table 1). Only one confirmed case (case 16) had documented evidence of age-appropriate vaccination. Two children had contraindications for vaccination (case 8, 11) and another two (case 9, 10) were due or overdue for their first dose of vaccine (Table 1). Of the 16 confirmed cases, 14 were identified as part of small epidemiological clusters (Fig. 1). Many were household based clusters with one hospital based cluster. Two cases acquired measles overseas (Table 1). Molecular analysis confirmed clustering based on epidemiological findings in all clusters (Fig. 1). 4.1. Laboratory investigations Time between the onset of the rash and a clinical request for both swab/urine and sera to be specifically tested for measles infection ranged from zero to thirteen days (mean 3.7/median 2 days and mean 3.4/median 2 days, respectively). Two cases had respiratory specimens collected during a febrile respiratory illness, but measles testing was requested on these stored specimens retrospectively. Most (13/16) of the MV-positive cases were positive by both rt-PCR and serology, while one was positive by rt-PCR alone (Table 2). IFA was positive in 6/11 and 6/13 urine and respiratory specimens, respectively from confirmed cases (the additional had indeterminate result or were not available – Tables 2 and 3). The rtPCRs were negative in all excluded cases where tested, whilst with three respiratory swabs and two urine IFAs were considered false positive based on other clinical, epidemiological and laboratory features (Table 2). All excluded cases had clinical histories that did not

satisfy the clinical case definition, and on the basis of the combination of laboratory data and discussion with treating clinicians were deemed exclusions. 4.2. Measles virus genotypes The genotypes detected during the investigation were D4 (1/16), D8 (1/16) and D9 (11/16) (Table 1). Two samples were negative for sequencing PCR. 4.3. Clinical features of some cases that developed complications A 30 year old man (case 4, Table 1) presented seven days following onset of measles rash with fever, ataxia, confusion, seizures and a Glasgow Coma Scale of five. His brain CT and MRI showed no focal encephalitis or demyelination. The CSF-PCR was negative for herpes simplex, varicella, cytomegalovirus and MV. He required

Table 3 Percentage of positive samples of different laboratory markers for confirmed measles cases. Disease marker

Positive samples %

Swab PCR Serum IgM Urine PCR Serum PCR Urine IFA Swab IFA

100 (13/13) 100 (15/15) 90.9 (10/11) 66.6 (6/9)a 54.5 (6/11) 46.2 (6/13)

a

Performed on sera available at SEALS serology.

172

J. Jayamaha et al. / Journal of Clinical Virology 54 (2012) 168–173

four days of ventilation in an intensive care unit and on discharge had mildly impaired memory and cognition. A 34 year old pregnant woman (case 12) presented during her 19th week of gestation with fevers, rigors, sweats, nausea, vomiting, productive cough and erythematous, pruritic rash. She and her unborn baby did not develop any obstetric complications and she continued well on review at the high risk pregnancy clinic. A 34 year old man (case 15) presented with diarrhoea, vomiting, maculopapular rash, fever and features of pneumonia. Bacteriological causes of pneumonia were excluded. He developed severe conjunctivitis requiring ophthalmic treatment. He had elevated serum alanine transaminase (980 IU/L) with common causes of viral hepatitis excluded using serology.

5. Discussion This MV outbreak affected mostly unvaccinated young adults between 20 and 35 years of age, similar to other recent Australian outbreaks.9–11 Two cases had travelled during the exposure period; one to the Philippines (case 14-genotype D9) and one to Italy (case 1-genotype D4) (Table 1) contracting common genotypes reported in these countries.12,13 However, in 3 of the 5 clusters, epidemiological links to a source case or importation could not be determined. Australia has a pattern of viral genotypes that is consistent with elimination of endemic virus. Genotypes associated with importation are D4, D8, H1, D9 and D5.14 Young adult travellers have been identified as a major source of imported infection in Australia and some European countries.10 Prevention of outbreaks in this susceptible age group may require strategies such as requirements for up-to-date vaccination records for entry into further education (case 13 – a university student) and prior to overseas travel.15 The single immunized case in this outbreak presented with a milder illness at age 19 years, consistent with previous reports.16 He visited several health care facilities and was in the emergency department for 2–3 h before a measles diagnosis was considered, resulting in the need for tracing of 60–70 contacts. Contacts were screened clinically and did not develop measles. Some other confirmed cases were also not identified until after they had visited doctors’ surgeries or emergency departments. This highlights the lack of experience of Australian clinicians in identifying measles, even when the presentation is consistent with the illness, as they may never have seen a case in this era of measles elimination. Vaccinating children at the appropriate age is important as two cases (case 9 and 10) who contracted measles (Table 1) were due or overdue for MMR vaccine. The current schedule of measles vaccine in Australia is at 12 months and 4 years of age. Coverage estimates in 2008 were 94.0% for the first dose and 80.3% for the second dose.17 Immunization histories for eight adult confirmed cases were not available (Table 1) as they did not have a personally held record and the Australian Childhood Immunization Register (ACIR), which is designed record vaccinations given to children aged up to and including 7 years, was not in existence when their childhood vaccinations were due. Two adult siblings were not vaccinated in their childhood as parents were conscientious objector. Parents may claim conscientious objection as grounds for not vaccinating their children which is recorded on ACIR. In 2008, 1.4% of children were entered in ACIR as conscientious objectors.17 Conscientious objection is not recorded for young adults. Four cases presented with a rash post vaccination; one (case 2) had known contact with her sibling (case 1) and three were deemed vaccine induced rashes and excluded. Employing other disease markers will help to differentiate mild measles infection and vaccination induced rash which otherwise would be difficult to differentiate only by serology.

This report shows the performance of different disease markers of measles in the setting of an outbreak (Tables 2 and 3). Measles virus is excreted for a longer period in urine and this may be used to determine the epidemiological link in late presentation.18 However respiratory swab/NPA PCR was the most sensitive test, detecting all cases where tested. This is likely due to the fact that all respiratory specimens were collected within eight days of rash when respiratory tract shedding would still be at significant levels (Table 3). However IFA has the fastest turnaround time per specimen and is often available in laboratories where PCR for measles is not available. IgM detection may be negative if serum is collected in the days immediately after rash onset and warrants testing on another sample collected during a subsequent visit. However if a single set of serum, respiratory and urine samples are collected on the same day, this combination will generally provide a rapid and sensitive diagnosis. Every attempt should be made in this regard so that a prompt and accurate diagnosis can be made to initiate a public health response. This reinforces the need for clinicians to notify public health authorities on suspicion of a case of measles; public health personnel may then facilitate the referral of all appropriate samples for the recommended range of assays. Real time PCR is highly sensitive, specific and can be applied on different sample types as seen in this outbreak. However it is often only available in reference or tertiary care laboratories and turnaround times may be relatively slow for public health activities. We did not attempt virus isolation as it would not aid rapid diagnosis required for infection control measures, and viral isolates were not required for genotyping since sequencing was performed directly on amplicons derived from clinical specimens by PCR. We have not done a formal cost analysis as measles investigations in Australia are relatively small in number compared with other investigations. Further, the cost of additional case ascertainment is significantly higher than testing, with accurate diagnosis (requiring nucleic acid testing) critical to directing such case finding. The additional sensitivity and specificity provided by nucleic acid testing (almost 100% compared with approximately 50% for IFA alone) means the additional cost of PCR ($AUS 40 for PCR, $AUS 15 for each serology assay) is a small proportion of the cost of case investigation. Complications of measles are well described elsewhere.19 Pregnancy is a high risk situation for measles.20 The pregnant female (case 12) and her fetus did not have adverse outcomes, although prematurity (31%), abortions (8%) and congenital malformations have been previously described. In addition, the incidence of death (3%) and other complications such as pneumonia (26%), from measles during pregnancy may be higher than expected for agecomparable, non-pregnant women.20 In conclusion, strategies should be reinforced to address the susceptibility of unvaccinated young adults and travellers to measles. Even in the measles elimination phase, clinicians still need to recognize and consider measles as a differential diagnosis when a patient presents with a fever and rash so that rapid detection and public health control measures can be instituted to prevent further transmission. Genotyping can be used as a powerful molecularepidemiological tool in outbreak investigation in countries that have reached elimination phase.

Funding None.

Competing interests None declared.

J. Jayamaha et al. / Journal of Clinical Virology 54 (2012) 168–173

Ethical approval Not required. Acknowledgements The authors thank staff of Wollongong and Randwick offices of the South Eastern Sydney Illawarra Public Health Unit who investigated each measles notification enabling us to describe the epidemiology of this outbreak, especially Sue Botham. Staff of Infectious Diseases, Infection Control, Paediatric and Emergency Departments of Wollongong, St. Vincent’s, and St. George’s hospitals, staff of SEALS virology, serology and VIDRL. All information presented in this paper was gathered as part of a series of public health investigations of identified individuals notified to the Public Health Unit with suspected or confirmed measles. These investigations were conducted in accordance with the NSW Health control guideline for measles and the Public Health Act 1991 (NSW) and no ethical approval was required. References 1. Heyman DL, editor. Control of communicable disease manual. 19th ed. Washington, DC: American Public Health Association; 2008. p. 120–4. 2. Centers for Disease Control and, Prevention. Increased transmission and outbreaks of measles – European region, 2011. MMWR 2011;60(47):1605–10. 3. National Notifiable Diseases Surveillance System. Notification Rate of Measles, received from State and Territory health authorities in the period of 1991 to 2011 and year-to-date notifications for 2012. http://www9.health. gov.au/cda/source/Rpt 4.cfm [accessed 15.01.12]. 4. Durrheim DN, Kelly H, Ferson MJ, Featherstone D. Remaining measles challenges in Australia. Med J Aust 2007;187:181–4. 5. World Health Organization. Manual for the laboratory diagnosis of measles viral infection—December 1999. Geneva: World Health Organization, Department of Vaccines and Biologicals; 2000.

173

6. National notifiable diseases case definitions; 2011. Available from: http://www.health.gov.au/casedefinitions [accessed 06.06.11]. 7. Smaron MF, Saxon E, Wood L, McCarthy C, Morello JA. Diagnosis of measles by fluorescent antibody and culture of nasopharyngeal secretions. J Virol Methods 1991;33:223–9. 8. Chibo D, Birch CJ, Rota PA, Catton MG. Molecular characterization of measles viruses isolated in Victoria, Australia, between 1973 and 1998. J Gen Virol 2000;81:2511–8. 9. Fielding JE. An outbreak of measles in Adelaide. Commun Dis Intell 2005;29:80–2. 10. Martin N, Foxwell AR. Measles status in Australia and outbreaks in the first quarter of 2009. Commun Dis Intell 2009;33:225–31. 11. Davidson N, Andrews R, Riddell M, Leydon J, Lynch P. A measles outbreak among young adults in Victoria, February 2001. Commun Dis Intell 2002;26:273–8. 12. World Health Organization. Report on the 8th WHO Global Measles Rubella laboratory network meeting; 2010. Available from http://www.who.int/ immunization delivery/adc/measles/Summary Recs LabNet 2010.pdf [accessed 06.06.11]. 13. Curtale F, Perrelli F, Mantovani J, Atti MC, Filia A, Nicoletti L, et al. Description of two measles outbreaks in the Lazio Region. Italy (2006–2007). Importance of pockets of low vaccine coverage in sustaining the infection. BMC Infect Dis 2010;10:1–9. 14. Rota P, Brown K, Mankertz A, Santibanez S, Shulga S, Muller C, et al. Global distribution of measles genotypes and measles molecular epidemiology. J Infect Dis 2011;204:S514–20. 15. Andrews N, Tischer A, Siedler A, Pebody RG, Barbara C, Cotter S, et al. Towards elimination: measles susceptibility in Australia and 17 European countries. Bull World Health Organ 2008;86:197–204. 16. Sheppeard V, Forssman B, Ferson MJ, Moreira C, Campbell-Lloyd S, Dwyer DE, et al. Vaccine failures and vaccine effectiveness in children during measles outbreak in New South Wales, March–May 2006. Commun Dis Intell 2009;33: 21–6. 17. Hull BP, Mahajan D, Dey A, Menzies RI, McIntyre PB. Immunization coverage annual report, 2008. Commun Dis Intell 2010;34:241–58. 18. van Binnendijk RS, van den Hof S, van den Kerkhof H, Kohl RHG, Woonink F, Berbers GAM, et al. Evaluation of serological and virological tests in the diagnosis of clinical and subclinical measles virus infections during an outbreak of measles in The Netherlands. J Infect Dis 2003;188:898–903. 19. Bellani WJ, Icenogle JP. Measles and rubella viruses. In: Murray PR, Baron EJ, Jorgensen JH, editors. Manual of clinical microbiology, vol. 2, 9th ed. Washington: American Society for Microbiology; 2007. p. 1378–91. 20. Eberhart-Phillips JE, Frederick PD, Baron RC, Mascola L. Measles in pregnancy: a descriptive study of 58 cases. Obstet Gynecol 1993;82:797–801.