‘Simple’ backache is the term used to describe back pain which is musculoskeletal in origin, mechanical in nature, varies with physical activity and time and can affect the lumbosacral region, buttocks and thighs in a person who is generally well. The lower back is the most commonly reported site of pain (MacArthur et al., 1991, Östgaard and Anderson, 1992, Russell et al., 1993 and Turgut et al., 1998).

Backache is very common during pregnancy, affecting 50% or more of all women to varying degrees (Berg et al., 1988, Fast et al., 1987 and Östgaard and Anderson, 1991). It is thought to be triggered by hormonal factors and the increased weight of the gravid uterus (MacLennan et al 1986). Several studies have also shown backache to be common following childbirth, with a prevalence range of 20–50% varying according to definition and timing. MacArthur et al (1991), in a study of 11,701 postnatal women questioned 1–9 years after the birth, found that 23% reported backache that had started within 3 months of giving birth and lasted for longer than 6 weeks, with 14% reporting this as a new symptom. Many more reported postpartum backache but could not accurately date these symptoms. Brown & Lumley (1998), in an Australian study of health problems among 1336 women at 6–7 months postpartum, found that backache was reported as a problem at some time since the birth by 44% of this sample.

Several longitudinal studies have found that postpartum backache is often not transient. Östgaard and Anderson, 1991 and Östgaard and Anderson, 1992, in a Swedish cohort study, followed a representative sample of 817 women through pregnancy, at the end of which 67% reported back pain and at 12–18 month follow-up back pain prevalence was 37%. A 20% sample (n = 1249) of all women who delivered during 1 year in one health region of Scotland were given a questionnaire about health problems in hospital, then a postal questionnaire at 8 weeks postpartum, half receiving another at 12–18 months, to investigate subsequent problems. Backache was reported by 22% of the women in hospital, by 24% between then and 8 weeks, and by 20% after this (Glazener et al 1993). Among 1042 women delivering in a maternity unit in Boston, USA, 44% reported backache at 1–2 months postpartum (Breen et al 1994). At 12–18 month follow-up, rates had not changed, with 49% experiencing back pain during the preceding 3 months (Groves et al 1994). At the time of follow-up, backache was more common (66%) among the women who had reported backache in the 1–2 month questionnaire, compared with those who had not (21%). A longitudinal study of the health of postpartum women in France (n = 589) and Italy (n = 697), followed at 5 and 12 months, found a high prevalence of backache in each country at both times: 49% and 50% in France, and 47% and 65% in Italy. The Italian rates showed a surprising increase over the time (Saurel-Cubizolles et al 2000).

A postal questionnaire survey for the Audit Commission (Garcia et al 1998), of a population-based sample of 2406 women throughout England and Wales at 4 months postpartum, asked about health problems as part of a wider study of maternity care. The women were asked to think back to 10 days, 1 month and 3 months postpartum and say which of a number of health problems they had at those times: 35% reported having had backache at 10 days, 27% at 1 month and 28% at 3 months.

A few studies have examined the severity of postpartum back pain. Bick & MacArthur (1995) investigated the severity and effect of various morbidities, including backache, in a sample of 1278 women by postal questionnaire at 6–7 months postpartum. There were 582 women (46%) who reported backache (new and recurrent) occurring within 3 months of birth and lasting for longer than 6 weeks; 49% of these women considered that it had affected their day-to-day activities. Mean severity of the backache, rated on a 100 mm visual analogue scale, was 39.4. Östgaard & Anderson (1992) found similar severity ratings of backache in their sample, with an average of 3.2 on a 10 cm visual analogue scale. Average pain severity before and during pregnancy in this same cohort (surveyed earlier using the same instruments) had been 0.99 and 4.4 respectively (Östgaard & Anderson 1991). Serious postpartum backache was reported in this study by 7% of women at 12–18 months, similar to the Boston study (Groves et al 1994), in which 8% reported severe back pain at this time.

Although backache is common following childbirth and can be persistent and affect daily life, studies have found medical consultation rates to be low (MacArthur et al., 1991, Bick and MacArthur, 1995 and Brown and Lumley, 1998). It should also be noted that general population-based epidemiological studies have found prevalence estimates of back pain to be high, 14–30% in studies asking about pain on that day and 30–40% in those asking about pain in the last month (CSAG 1995). It is difficult to assess the extent of additional back pain that is attributable to childbirth.

A previous history of back pain is an important risk factor for postnatal backache and some studies have also found a relationship with physically demanding work. Östgaard & Anderson (1992), in the cohort study described earlier, found a significant association with back pain before pregnancy; sick leave for back pain during pregnancy; and physically heavy work (the researchers were not able to determine if the effect was from work before, during or after the pregnancy). Breen et al (1994), in the Boston study, also found a history of back pain to be predictive of postpartum back pain, but only if there had also been back pain during pregnancy: Turgut et al (1998), in Turkey, followed 88 women who had back pain during pregnancy to 6 months postpartum and found that a history of pre-pregnancy back pain was a significant predictor of pain at 6 months. They found no relationship with heavy work before pregnancy but this was based on small numbers.

Breen et al (1994), in the Boston study, found younger maternal age and greater maternal weight to be predictors of postpartum backache, and Brown & Lumley (1998) found an association with heavier infant birthweight. Glazener et al (1995) found a significant association with mode of delivery, backache more likely after instrumental and caesarean deliveries. Brown & Lumley (1998) found proportionally more backache reported as a problem following these types of births but the difference was not statistically significant. MacArthur et al (1991) found an association with ethnic group, Asian women being much more likely to report backache as well as other musculoskeletal symptoms, although this may be due to cultural differences in the reporting of morbidity (MacArthur et al 1993). Loughnan et al (2002), based on data from a randomised controlled trial (RCT) of epidural versus meperidine for labour analgesia (see below), found that being non-Caucasian was an independent predictor of new backache at 6 months postpartum. Exact ethnic group was not recorded but the authors state that the hospital catchment population is 25% Asian.

The possibility of an association between epidural analgesia and postpartum backache has been investigated in several observational studies, some finding an association (MacArthur et al., 1990, Russell et al., 1993, MacLeod et al., 1995 and Brown and Lumley, 1998), others not (Breen et al., 1994, Macarthur et al., 1995 and Russell et al., 1996). The first study on this found that 18.9% of women reported new backache occurring within 3 months of delivery and lasting for over 6 weeks following epidural for analgesia, compared with 10.5% of those without (MacArthur et al 1990). The suggested mechanism was through stressed postures in labour, affected by the hormone relaxin and exacerbated when discomfort feedback is inhibited by epidural block. Russell et al (1993), in a similar study at St Thomas’ Hospital, London, of 612 primiparous women who received an epidural and 403 who did not, found a similar size epidural excess of new backache (17.8% vs 11.7%).

In the Boston study (Breen et al 1994) of backache at 1–2 months postpartum, epidural use was not associated with backache, either generally or with new symptoms. In a small Canadian study (Macarthur et al 1995), 164 women who had an epidural and 165 who had not were interviewed at 1 and 7 days and 6 weeks after delivery. After excluding women who had had pregnancy backache, the only difference in new postpartum backache that reached statistical significance was on day 1 (52% epidural vs 39% non-epidural), although there was a non-significant twofold epidural excess at 6 weeks (15% vs 7%). Among those followed up at 1 year, 10% had backache in the epidural and 14% in the non-epidural groups (Macarthur et al 1997).

A second study from St Thomas’ Hospital, London (Russell et al 1996), compared women who requested an epidural, randomised to either a traditional (n = 157) or a mobile technique (n = 162), and a third group (n = 131) with no epidural, recruited by taking the next parity-matched delivery in the birth register. A postal questionnaire at 3 months postpartum found backache reported by 39% of the traditional group, 30% of the mobile group and 30.5% of the group with no epidural, and new backache was reported by 6.4%, 8.6% and 6.9% respectively. Like the Canadian study, since the differences were not statistically significant, the authors concluded that women could be reassured that epidurals are not associated with postpartum backache. However, both studies were small with insufficient power to detect the size of difference found in the earlier studies.

In general, the observational studies which showed no backache excess were in hospitals using a lower concentration of local anaesthetic together with an opiate, which produces less dense motor block, often allowing ambulation. A recent RCT of traditional versus two types of low-dose ‘mobile’ epidural techniques assessed long-term backache as an outcome (COMET Study Group 2003). This showed no significant difference in backache starting within 6 weeks of the birth and lasting longer than 3 months in either of the mobile techniques, although there was proportionally less backache in the combined-spinal than the traditional epidural group (43% vs 50%).

There have now been numerous RCTs designed to examine various possible effects of epidural versus no epidural labours, two of which have assessed backache as an outcome measure (Amin-Samuah et al 2005). Loughnan et al., 2000 and Loughnan et al., 2002 randomised 611 nulliparous women, of whom 249 were allocated to receive meperidine and 259 to receive epidural, and returned a 6-month questionnaire about long-term backache. The intention-to-treat comparisons showed no significant difference in any backache present at 6 months (prevalence 48% and 50% respectively) or in new backache (Loughnan et al 2002). However, there was substantial treatment cross-over, with 57% of those randomised to the meperidine group actually having an epidural, and 15% of those randomised to epidural not having one (Loughnan et al 2000). In the smaller trial by Howell et al., 2001 and Howell et al., 2002 184 nulliparous women were randomised to epidural and 184 to no epidural, of whom 151 and 155 respectively were followed to a median of 26 months postpartum. Again, no significant differences in backache between trial groups were found. Cross-over in this trial was less, but still 28% of those randomised to no epidural had one and 33% of those randomised to epidural did not have one.

As shown above, most of the studies investigating the relationship between epidurals and backache have been observational, and since we know that women who receive epidurals generally have less straightforward labours and deliveries, these differences may account for the association. Findings from well-conducted RCTs can generally provide conclusive evidence but because epidurals provide superior pain relief than other methods, many of the women randomised to receive no epidural are likely to end up having one. It is more difficult to interpret results of RCTs where there is substantial treatment cross-over, since this will dilute any possible effect of an intervention, although a large trial would still show a large difference. The likely conclusion, therefore, on whether epidural is a risk factor for longer term postpartum backache is that if it is, any effect is likely to be relatively small.